1CONDOR_SUBMIT(1) HTCondor Manual CONDOR_SUBMIT(1)
2
6 condor_submit - HTCondor Manual
7 Queue jobs for execution under HTCondor
9
12 condor_submit [-terse ] [-verbose ] [-unused ] [-file submit_file]
13 [-name schedd_name] [-remote schedd_name] [-addr <ip:port>] [-pool
14 pool_name] [-disable ] [-password passphrase] [-debug ] [-append com‐
15 mand ...][-batch-name batch_name] [-spool ] [-dump filename] [-interac‐
16 tive ] [-allow-crlf-script ] [-dry-run ] [-maxjobs number-of-jobs]
17 [-single-cluster ] [-stm method] [<submit-variable>=<value> ] [submit
18 description file ] [-queue queue_arguments]
19
21 condor_submit is the program for submitting jobs for execution under
22 HTCondor. condor_submit requires one or more submit description com‐
23 mands to direct the queuing of jobs. These commands may come from a
24 file, standard input, the command line, or from some combination of
25 these. One submit description may contain specifications for the queu‐
26 ing of many HTCondor jobs at once. A single invocation of condor_submit
27 may cause one or more clusters. A cluster is a set of jobs specified in
28 the submit description between queue commands for which the exe‐
29 cutable is not changed. It is advantageous to submit multiple jobs as a
30 single cluster because:
31 • Much less memory is used by the scheduler to hold the same number of
33 jobs.
34 • Only one copy of the checkpoint file is needed to represent all jobs
36 in a cluster until they begin execution.
37 • There is much less overhead involved for HTCondor to start the next
39 job in a cluster than for HTCondor to start a new cluster. This can
40 make a big difference when submitting lots of short jobs.
41 Multiple clusters may be specified within a single submit description.
43 Each cluster must specify a single executable.
44 The job ClassAd attribute ClusterId identifies a cluster.
46 The submit description file argument is the path and file name of the
48 submit description file. If this optional argument is the dash charac‐
49 ter (-), then the commands are taken from standard input. If - is spec‐
50 ified for the submit description file, -verbose is implied; this can be
51 overridden by specifying -terse.
52 If no submit discription file argument is given, and no -queue argument
54 is given, commands are taken automatically from standard input.
55 Note that submission of jobs from a Windows machine requires a stashed
57 password to allow HTCondor to impersonate the user submitting the job.
58 To stash a password, use the condor_store_cred command. See the manual
59 page for details.
60 For lengthy lines within the submit description file, the backslash (\)
62 is a line continuation character. Placing the backslash at the end of a
63 line causes the current line's command to be continued with the next
64 line of the file. Submit description files may contain comments. A com‐
65 ment is any line beginning with a pound character (#).
66
68 -terse Terse output - display JobId ranges only.
69 -verbose
71 Verbose output - display the created job ClassAd
72 -unused
74 As a default, causes no warnings to be issued about user-de‐
75 fined macros not being used within the submit description
76 file. The meaning reverses (toggles) when the configuration
77 variable WARN_ON_UNUSED_SUBMIT_FILE_MACROS
78 is set to the non default value of False. Printing the
79 warnings can help identify spelling errors of submit descrip‐
80 tion file commands. The warnings are sent to stderr.
81 -file submit_file
83 Use submit_file as the submit discription file. This is
84 equivalent to providing submit_file as an argument without
85 the preceeding -file.
86 -name schedd_name
88 Submit to the specified condor_schedd. Use this option to
89 submit to a condor_schedd other than the default local one.
90 schedd_name is the value of the Name ClassAd attribute on the
91 machine where the condor_schedd daemon runs.
92 -remote schedd_name
94 Submit to the specified condor_schedd, spooling all required
95 input files over the network connection. schedd_name is the
96 value of the Name ClassAd attribute on the machine where the
97 condor_schedd daemon runs. This option is equivalent to using
98 both -name and -spool.
99 -addr <ip:port>
101 Submit to the condor_schedd at the IP address and port given
102 by the sinful string argument <ip:port>.
103 -pool pool_name
105 Look in the specified pool for the condor_schedd to submit
106 to. This option is used with -name or -remote.
107 -disable
109 Disable file permission checks when submitting a job for read
110 permissions on all input files, such as those defined by com‐
111 mands input and transfer_input_files , as well as write
112 permission to output files, such as a log file defined by log
113 and output files defined with output or transfer_out‐
114 put_files .
115 -password passphrase
117 Specify a password to the MyProxy server.
118 -debug Cause debugging information to be sent to stderr, based on
120 the value of the configuration variable TOOL_DEBUG.
121 -append command
123 Augment the commands in the submit description file with the
124 given command. This command will be considered to immediately
125 precede the queue command within the submit description file,
126 and come after all other previous commands. If the command
127 specifies a queue command, as in the example
128 condor_submit mysubmitfile -append "queue input in A, B, C"
130 then the entire -append command line option and its arguments
132 are converted to
133 condor_submit mysubmitfile -queue input in A, B, C
135 The submit description file is not modified. Multiple com‐
137 mands are specified by using the -append option multiple
138 times. Each new command is given in a separate -append op‐
139 tion. Commands with spaces in them will need to be enclosed
140 in double quote marks.
141 -batch-name batch_name
143 Set the batch name for this submit. The batch name is dis‐
144 played by condor_q -batch. It is intended for use by users to
145 give meaningful names to their jobs and to influence how con‐
146 dor_q groups jobs for display. Use of this argument takes
147 precedence over a batch name specified in the submit descrip‐
148 tion file itself.
149 -spool Spool all required input files, job event log, and proxy over
151 the connection to the condor_schedd. After submission, modify
152 local copies of the files without affecting your jobs. Any
153 output files for completed jobs need to be retrieved with
154 condor_transfer_data.
155 -dump filename
157 Sends all ClassAds to the specified file, instead of to the
158 condor_schedd.
159 -interactive
161 Indicates that the user wants to run an interactive shell on
162 an execute machine in the pool. This is equivalent to creat‐
163 ing a submit description file of a vanilla universe sleep
164 job, and then running condor_ssh_to_job by hand. Without any
165 additional arguments, condor_submit with the -interactive
166 flag creates a dummy vanilla universe job that sleeps, sub‐
167 mits it to the local scheduler, waits for the job to run, and
168 then launches condor_ssh_to_job to run a shell. If the user
169 would like to run the shell on a machine that matches a par‐
170 ticular requirements expression, the submit description file
171 is specified, and it will contain the expression. Note that
172 all policy expressions specified in the submit description
173 file are honored, but any executable or universe commands
174 are overwritten to be sleep and vanilla. The job ClassAd at‐
175 tribute InteractiveJob is set to True to identify interactive
176 jobs for condor_startd policy usage.
177 -allow-crlf-script
179 Changes the check for an invalid line ending on the exe‐
180 cutable script's #! line from an ERROR to a WARNING. The #!
181 line will be ignored by Windows, so it won't matter if it is
182 invalid; but Unix and Linux will not run a script that has a
183 Windows/DOS line ending on the first line of the script. So
184 condor_submit will not allow such a script to be submitted as
185 the job's executable unless this option is supplied.
186 -dry-run file
188 Parse the submit description file, sending the resulting job
189 ClassAd to the file given by file, but do not submit the
190 job(s). This permits observation of the job specification,
191 and it facilitates debugging the submit description file con‐
192 tents. If file is -, the output is written to stdout.
193 -maxjobs number-of-jobs
195 If the total number of jobs specified by the submit descrip‐
196 tion file is more than the integer value given by num‐
197 ber-of-jobs, then no jobs are submitted for execution and an
198 error message is generated. A 0 or negative value for the
199 number-of-jobs causes no limit to be imposed.
200 -single-cluster
202 If the jobs specified by the submit description file causes
203 more than a single cluster value to be assigned, then no jobs
204 are submitted for execution and an error message is gener‐
205 ated.
206 -stm method
208 Specify the method use to move a sandbox into HTCondor.
209 method is one of stm_use_schedd_only or stm_use_transferd.
210 <submit-variable>=<value>
212 Defines a submit command or submit variable with a value, and
213 parses it as if it was placed at the beginning of the submit
214 description file. The submit description file is not changed.
215 To correctly parse the condor_submit command line, this op‐
216 tion must be specified without white space characters before
217 and after the equals sign (=), or the entire option must be
218 surrounded by double quote marks.
219 -queue queue_arguments
221 A command line specification of how many jobs to queue, which
222 is only permitted if the submit description file does not
223 have a queue command. The queue_arguments are the same as may
224 be within a submit description file. The parsing of the
225 queue_arguments finishes at the end of the line or when a
226 dash character (-) is encountered. Therefore, its best place‐
227 ment within the command line will be at the end of the com‐
228 mand line.
229 On a Unix command line, the shell expands file globs before
231 parsing occurs.
232
234 Note: more information on submitting HTCondor jobs can be found here:
235 /users-manual/submitting-a-job.
236 As of version 8.5.6, the condor_submit language supports multi-line
238 values in commands. The syntax is the same as the configuration lan‐
239 guage (see more details here: admin-manual/introduction-to-configura‐
240 tion:multi-line values).
241 Each submit description file describes one or more clusters of jobs to
243 be placed in the HTCondor execution pool. All jobs in a cluster must
244 share the same executable, but they may have different input and output
245 files, and different program arguments. The submit description file is
246 generally the last command-line argument to condor_submit. If the sub‐
247 mit description file argument is omitted, condor_submit will read the
248 submit description from standard input.
249 The submit description file must contain at least one executable com‐
251 mand and at least one queue command. All of the other commands have de‐
252 fault actions.
253 Note that a submit file that contains more than one executable command
255 will produce multiple clusters when submitted. This is not generally
256 recommended, and is not allowed for submit files that are run as DAG
257 node jobs by condor_dagman.
258 The commands which can appear in the submit description file are numer‐
260 ous. They are listed here in alphabetical order by category.
261 BASIC COMMANDS
263 arguments = <argument_list>
265 List of arguments to be supplied to the executable as part of
266 the command line.
267 In the java universe, the first argument must be the name of
269 the class containing main.
270 There are two permissible formats for specifying arguments,
272 identified as the old syntax and the new syntax. The old syn‐
273 tax supports white space characters within arguments only in
274 special circumstances; when used, the command line arguments
275 are represented in the job ClassAd attribute Args. The new
276 syntax supports uniform quoting of white space characters
277 within arguments; when used, the command line arguments are
278 represented in the job ClassAd attribute Arguments.
279 Old Syntax
281 In the old syntax, individual command line arguments are de‐
283 limited (separated) by space characters. To allow a double
284 quote mark in an argument, it is escaped with a backslash;
285 that is, the two character sequence \" becomes a single dou‐
286 ble quote mark within an argument.
287 Further interpretation of the argument string differs depend‐
289 ing on the operating system. On Windows, the entire argument
290 string is passed verbatim (other than the backslash in front
291 of double quote marks) to the Windows application. Most Win‐
292 dows applications will allow spaces within an argument value
293 by surrounding the argument with double quotes marks. In all
294 other cases, there is no further interpretation of the argu‐
295 ments.
296 Example:
298 arguments = one \"two\" 'three'
300 Produces in Unix vanilla universe:
302 argument 1: one
304 argument 2: "two"
305 argument 3: 'three'
306 New Syntax
308 Here are the rules for using the new syntax:
310 1. The entire string representing the command line arguments
312 is surrounded by double quote marks. This permits the
313 white space characters of spaces and tabs to potentially
314 be embedded within a single argument. Putting the double
315 quote mark within the arguments is accomplished by escap‐
316 ing it with another double quote mark.
317 2. The white space characters of spaces or tabs delimit argu‐
319 ments.
320 3. To embed white space characters of spaces or tabs within a
322 single argument, surround the entire argument with single
323 quote marks.
324 4. To insert a literal single quote mark, escape it within an
326 argument already delimited by single quote marks by adding
327 another single quote mark.
328 Example:
330 arguments = "3 simple arguments"
332 Produces:
334 argument 1: 3
336 argument 2: simple
337 argument 3: arguments
338 Another example:
340 arguments = "one 'two with spaces' 3"
342 Produces:
344 argument 1: one
346 argument 2: two with spaces
347 argument 3: 3
348 And yet another example:
350 arguments = "one ""two"" 'spacey ''quoted'' argument'"
352 Produces:
354 argument 1: one
356 argument 2: "two"
357 argument 3: spacey 'quoted' argument
358 Notice that in the new syntax, the backslash has no special
360 meaning. This is for the convenience of Windows users.
361 environment = <parameter_list>
364 List of environment
365 variables.
366 There are two different formats for specifying the environ‐
368 ment variables: the old format and the new format. The old
369 format is retained for backward-compatibility. It suffers
370 from a platform-dependent syntax and the inability to insert
371 some special characters into the environment.
372 The new syntax for specifying environment values:
374 1. Put double quote marks around the entire argument string.
376 This distinguishes the new syntax from the old. The old
377 syntax does not have double quote marks around it. Any
378 literal double quote marks within the string must be es‐
379 caped by repeating the double quote mark.
380 2. Each environment entry has the form
382 <name>=<value>
384 3. Use white space (space or tab characters) to separate en‐
386 vironment entries.
387 4. To put any white space in an environment entry, surround
389 the space and as much of the surrounding entry as desired
390 with single quote marks.
391 5. To insert a literal single quote mark, repeat the single
393 quote mark anywhere inside of a section surrounded by sin‐
394 gle quote marks.
395 Example:
397 environment = "one=1 two=""2"" three='spacey ''quoted'' value'"
399 Produces the following environment entries:
401 one=1
403 two="2"
404 three=spacey 'quoted' value
405 Under the old syntax, there are no double quote marks sur‐
407 rounding the environment specification. Each environment en‐
408 try remains of the form
409 <name>=<value>
411 Under Unix, list multiple environment entries by separating
413 them with a semicolon (;). Under Windows, separate multiple
414 entries with a vertical bar (|). There is no way to insert a
415 literal semicolon under Unix or a literal vertical bar under
416 Windows. Note that spaces are accepted, but rarely desired,
417 characters within parameter names and values, because they
418 are treated as literal characters, not separators or ignored
419 white space. Place spaces within the parameter list only if
420 required.
421 A Unix example:
423 environment = one=1;two=2;three="quotes have no 'special' meaning"
425 This produces the following:
427 one=1
429 two=2
430 three="quotes have no 'special' meaning"
431 If the environment is set with the environment command and
433 getenv is also set to true, values specified with environ‐
434 ment override values in the submitter's environment (regard‐
435 less of the order of the environment and getenv commands).
436 error = <pathname>
439 A path and file name used by HTCondor to capture any error
440 messages the program would normally write to the screen (that
441 is, this file becomes stderr). A path is given with respect
442 to the file system of the machine on which the job is submit‐
443 ted. The file is written (by the job) in the remote scratch
444 directory of the machine where the job is executed. When the
445 job exits, the resulting file is transferred back to the ma‐
446 chine where the job was submitted, and the path is utilized
447 for file placement. If not specified, the default value of
448 /dev/null is used for submission to a Unix machine. If not
449 specified, error messages are ignored for submission to a
450 Windows machine. More than one job should not use the same
451 error file, since this will cause one job to overwrite the
452 errors of another. If HTCondor detects that the error and
453 output files for a job are the same, it will run the job such
454 that the output and error data is merged.
455 executable = <pathname>
457 An optional path and a required file name of the executable
458 file for this job cluster. Only one executable command
459 within a submit description file is guaranteed to work prop‐
460 erly. More than one often works.
461 If no path or a relative path is used, then the executable
463 file is presumed to be relative to the current working direc‐
464 tory of the user as the condor_submit command is issued.
465 If submitting into the standard universe, then the named exe‐
467 cutable must have been re-linked with the HTCondor libraries
468 (such as via the condor_compile command). If submitting into
469 the vanilla universe (the default), then the named executable
470 need not be re-linked and can be any process which can run in
471 the background (shell scripts work fine as well). If submit‐
472 ting into the Java universe, then the argument must be a com‐
473 piled .class file.
474 getenv = <True | False>
477 If getenv is set to
478 True, then condor_submit will copy all of the user's current
479 shell environment variables at the time of job submission
480 into the job ClassAd. The job will therefore execute with the
481 same set of environment variables that the user had at submit
482 time. Defaults to False.
483 If the environment is set with the environment command and
485 getenv is also set to true, values specified with environment
486 override values in the submitter's environment (regardless of
487 the order of the environment and getenv commands).
488 input = <pathname>
490 HTCondor assumes that its jobs are long-running, and that the
491 user will not wait at the terminal for their completion. Be‐
492 cause of this, the standard files which normally access the
493 terminal, (stdin, stdout, and stderr), must refer to files.
494 Thus, the file name specified with input should contain any
495 keyboard input the program requires (that is, this file be‐
496 comes stdin). A path is given with respect to the file system
497 of the machine on which the job is submitted. The file is
498 transferred before execution to the remote scratch directory
499 of the machine where the job is executed. If not specified,
500 the default value of /dev/null is used for submission to a
501 Unix machine. If not specified, input is ignored for submis‐
502 sion to a Windows machine. For grid universe jobs, input
503 may be a URL that the Globus tool globus_url_copy under‐
504 stands.
505 Note that this command does not refer to the command-line ar‐
507 guments of the program. The command-line arguments are speci‐
508 fied by the arguments command.
509 log = <pathname>
512 Use log to specify a file name where HTCondor will write a
513 log file of what is happening with this job cluster, called a
514 job event log. For example, HTCondor will place a log entry
515 into this file when and where the job begins running, when
516 the job produces a checkpoint, or moves (migrates) to another
517 machine, and when the job completes. Most users find specify‐
518 ing a log file to be handy; its use is recommended. If no log
519 entry is specified, HTCondor does not create a log for this
520 cluster. If a relative path is specified, it is relative to
521 the current working directory as the job is submitted or the
522 directory specified by submit command initialdir on the sub‐
523 mit machine.
524 log_xml = <True | False>
527 If log_xml is True, then the job event log file will be
528 written in ClassAd XML. If not specified, XML is not used.
529 Note that the file is an XML fragment; it is missing the file
530 header and footer. Do not mix XML and non-XML within a single
531 file. If multiple jobs write to a single job event log file,
532 ensure that all of the jobs specify this option in the same
533 way.
534 notification = <Always | Complete | Error | Never>
539 Owners of HTCondor jobs are notified by e-mail when certain
540 events occur. If defined by Always, the owner will be noti‐
541 fied whenever the job produces a checkpoint, as well as when
542 the job completes. If defined by Complete, the owner will be
543 notified when the job terminates. If defined by Error, the
544 owner will only be notified if the job terminates abnormally,
545 (as defined by JobSuccessExitCode, if defined) or if the job
546 is placed on hold because of a failure, and not by user re‐
547 quest. If defined by Never (the default), the owner will not
548 receive e-mail, regardless to what happens to the job. The
549 HTCondor User's manual documents statistics included in the
550 e-mail.
551 notify_user = <email-address>
553 Used to specify the e-mail address to use when HTCondor sends
554 e-mail about a job. If not specified, HTCondor defaults to
555 using the e-mail address defined by
556 job-owner@UID_DOMAIN
558 where the configuration variable UID_DOMAIN
560 is specified by the HTCondor site administrator. If UID_DO‐
561 MAIN has not been specified, HTCondor sends the e-mail to:
562 job-owner@submit-machine-name
564 output = <pathname>
568 The output file captures any information the program would
569 ordinarily write to the screen (that is, this file becomes
570 stdout). A path is given with respect to the file system of
571 the machine on which the job is submitted. The file is writ‐
572 ten (by the job) in the remote scratch directory of the ma‐
573 chine where the job is executed. When the job exits, the re‐
574 sulting file is transferred back to the machine where the job
575 was submitted, and the path is utilized for file placement.
576 If not specified, the default value of /dev/null is used for
577 submission to a Unix machine. If not specified, output is ig‐
578 nored for submission to a Windows machine. Multiple jobs
579 should not use the same output file, since this will cause
580 one job to overwrite the output of another. If HTCondor de‐
581 tects that the error and output files for a job are the same,
582 it will run the job such that the output and error data is
583 merged.
584 Note that if a program explicitly opens and writes to a file,
586 that file should not be specified as the output file.
587 priority = <integer>
590 An HTCondor job priority can be any integer, with 0 being the
591 default. Jobs with higher numerical priority will run before
592 jobs with lower numerical priority. Note that this priority
593 is on a per user basis. One user with many jobs may use this
594 command to order his/her own jobs, and this will have no ef‐
595 fect on whether or not these jobs will run ahead of another
596 user's jobs.
597 Note that the priority setting in an HTCondor submit file
599 will be overridden by condor_dagman if the submit file is
600 used for a node in a DAG, and the priority of the node within
601 the DAG is non-zero (see users-manual/dagman-applica‐
602 tions:advanced features of dagman for more details).
603 queue [<int expr> ]
605 Places zero or more copies of the job into the HTCondor
606 queue.
607 queue [<int expr> ] [<varname> ] in [slice ] <list of items> Places
609 zero or more copies of the job in the queue based on items in
610 a <list of items>
611 queue [<int expr> ] [<varname> ] matching [files | dirs ] [slice ]
613 <list of items with file globbing>] Places zero or more
614 copies of the job in the queue based on files that match a
615 <list of items with file globbing>
616 queue [<int expr> ] [<list of varnames> ] from [slice ] <file name>
618 | <list of items>] Places zero or more copies of the job in
619 the queue based on lines from the submit file or from <file
620 name>
621 The optional argument <int expr> specifies how many times to
623 repeat the job submission for a given set of arguments. It
624 may be an integer or an expression that evaluates to an inte‐
625 ger, and it defaults to 1. All but the first form of this
626 command are various ways of specifying a list of items. When
627 these forms are used <int expr> jobs will be queued for each
628 item in the list. The in, matching and from keyword indicates
629 how the list will be specified.
630 • in The list of items is an explicit comma and/or space sep‐
632 arated <list of items>. If the <list of items> begins with
633 an open paren, and the close paren is not on the same line
634 as the open, then the list continues until a line that be‐
635 gins with a close paren is read from the submit file.
636 • matching Each item in the <list of items with file glob‐
638 bing> will be matched against the names of files and direc‐
639 tories relative to the current directory, the set of match‐
640 ing names is the resulting list of items.
641 • files Only filenames will matched.
643 • dirs Only directory names will be matched.
645 • from <file name> | <list of items> Each line from <file
647 name> or <list of items> is a single item, this allows for
648 multiple variables to be set for each item. Lines from
649 <file name> or <list of items> will be split on comma
650 and/or space until there are values for each of the vari‐
651 ables specified in <list of varnames>. The last variable
652 will contain the remainder of the line. When the <list of
653 items> form is used, the list continues until the first
654 line that begins with a close paren, and lines beginning
655 with pound sign ('#') will be skipped. When using the
656 <file name> form, if the <file name> ends with |, then it
657 will be executed as a script whatever the script writes to
658 stdout will be the list of items.
659 The optional argument <varname> or <list of varnames> is the
661 name or names of of variables that will be set to the value
662 of the current item when queuing the job. If no <varname> is
663 specified the variable ITEM will be used. Leading and trail‐
664 ing whitespace be trimmed. The optional argument <slice> is a
665 python style slice selecting only some of the items in the
666 list of items. Negative step values are not supported.
667 A submit file may contain more than one queue statement,
669 and if desired, any commands may be placed between subsequent
670 queue commands, such as new input , output , error ,
671 initialdir , or arguments commands. This is handy when
672 submitting multiple runs into one cluster with one submit de‐
673 scription file.
674 universe = <vanilla | standard | scheduler | local | grid | java| vm
677 | parallel | docker>
678 Specifies which HTCondor universe to use when running this
679 job. The HTCondor universe specifies an HTCondor execution
680 environment.
681 The vanilla universe is the default (except where the config‐
683 uration variable DEFAULT_UNIVERSE
684 defines it otherwise), and is an execution environment for
685 jobs which do not use HTCondor's mechanisms for taking check‐
686 points; these are ones that have not been linked with the HT‐
687 Condor libraries. Use the vanilla universe to submit shell
688 scripts to HTCondor.
689 The standard universe tells HTCondor that this job has been
691 re-linked via condor_compile with the HTCondor libraries and
692 therefore supports taking checkpoints and remote system
693 calls.
694 The scheduler universe is for a job that is to run on the ma‐
696 chine where the job is submitted. This universe is intended
697 for a job that acts as a metascheduler and will not be pre‐
698 empted.
699 The local universe is for a job that is to run on the machine
701 where the job is submitted. This universe runs the job imme‐
702 diately and will not preempt the job.
703 The grid universe forwards the job to an external job manage‐
705 ment system. Further specification of the grid universe is
706 done with the grid_resource command.
707 The java universe is for programs written to the Java Virtual
709 Machine.
710 The vm universe facilitates the execution of a virtual ma‐
712 chine.
713 The parallel universe is for parallel jobs (e.g. MPI) that
715 require multiple machines in order to run.
716 The docker universe runs a docker container as an HTCondor
718 job.
719 COMMANDS FOR MATCHMAKING
721 rank = <ClassAd Float Expression>
723 A ClassAd Floating-Point expression that states how to rank
724 machines which have already met the requirements expression.
725 Essentially, rank expresses preference. A higher numeric
726 value equals better rank. HTCondor will give the job the ma‐
727 chine with the highest rank. For example,
728 request_memory = max({60, Target.TotalSlotMemory})
730 rank = Memory
731 asks HTCondor to find all available machines with more than
733 60 megabytes of memory and give to the job the machine with
734 the most amount of memory. The HTCondor User's Manual con‐
735 tains complete information on the syntax and available at‐
736 tributes that can be used in the ClassAd expression.
737 request_cpus = <num-cpus>
740 A requested number of CPUs (cores). If not specified, the
741 number requested will be 1. If specified, the expression
742 && (RequestCpus <= Target.Cpus)
744 is appended to the requirements expression for the job.
746 For pools that enable dynamic condor_startd provisioning,
748 specifies the minimum number of CPUs requested for this job,
749 resulting in a dynamic slot being created with this many
750 cores.
751 request_disk = <quantity>
754 The requested amount of disk space in KiB requested for this
755 job. If not specified, it will be set to the job ClassAd at‐
756 tribute DiskUsage. The expression
757 && (RequestDisk <= Target.Disk)
759 is appended to the requirements expression for the job.
761 For pools that enable dynamic condor_startd provisioning, a
763 dynamic slot will be created with at least this much disk
764 space.
765 Characters may be appended to a numerical value to indicate
767 units. K or KB indicates KiB, 210 numbers of bytes. M or MB
768 indicates MiB, 220 numbers of bytes. G or GB indicates GiB,
769 230 numbers of bytes. T or TB indicates TiB, 240 numbers of
770 bytes.
771 request_memory = <quantity>
774 The required amount of memory in MiB that this job needs to
775 avoid excessive swapping. If not specified and the submit
776 command vm_memory is specified, then the value specified
777 for vm_memory defines request_memory . If neither re‐
778 quest_memory nor vm_memory is specified, the value is set
779 by the configuration variable JOB_DEFAULT_REQUESTMEMORY
780 . The actual amount of memory used by a job is represented
781 by the job ClassAd attribute MemoryUsage.
782 For pools that enable dynamic condor_startd provisioning, a
784 dynamic slot will be created with at least this much RAM.
785 The expression
787 && (RequestMemory <= Target.Memory)
789 is appended to the requirements expression for the job.
791 Characters may be appended to a numerical value to indicate
793 units. K or KB indicates KiB, 210 numbers of bytes. M or MB
794 indicates MiB, 220 numbers of bytes. G or GB indicates GiB,
795 230 numbers of bytes. T or TB indicates TiB, 240 numbers of
796 bytes.
797 request_<name> = <quantity>
802 The required amount of the custom machine resource identified
803 by <name> that this job needs. The custom machine resource is
804 defined in the machine's configuration. Machines that have
805 available GPUs will define <name> to be GPUs.
806 requirements = <ClassAd Boolean Expression>
809 The requirements command is a boolean ClassAd expression
810 which uses C-like operators. In order for any job in this
811 cluster to run on a given machine, this requirements expres‐
812 sion must evaluate to true on the given machine.
813 For scheduler and local universe jobs, the requirements ex‐
815 pression is evaluated against the Scheduler ClassAd which
816 represents the the condor_schedd daemon running on the submit
817 machine, rather than a remote machine. Like all commands in
818 the submit description file, if multiple requirements com‐
819 mands are present, all but the last one are ignored. By de‐
820 fault, condor_submit appends the following clauses to the re‐
821 quirements expression:
822 1. Arch and OpSys are set equal to the Arch and OpSys of the
824 submit machine. In other words: unless you request other‐
825 wise, HTCondor will give your job machines with the same
826 architecture and operating system version as the machine
827 running condor_submit.
828 2. Cpus >= RequestCpus, if the job ClassAd attribute Re‐
830 questCpus is defined.
831 3. Disk >= RequestDisk, if the job ClassAd attribute Request‐
833 Disk is defined. Otherwise, Disk >= DiskUsage is appended
834 to the requirements. The DiskUsage attribute is initial‐
835 ized to the size of the executable plus the size of any
836 files specified in a transfer_input_files command. It ex‐
837 ists to ensure there is enough disk space on the target
838 machine for HTCondor to copy over both the executable and
839 needed input files. The DiskUsage attribute represents the
840 maximum amount of total disk space required by the job in
841 kilobytes. HTCondor automatically updates the DiskUsage
842 attribute approximately every 20 minutes while the job
843 runs with the amount of space being used by the job on the
844 execute machine.
845 4. Memory >= RequestMemory, if the job ClassAd attribute Re‐
847 questMemory is defined.
848 5. If Universe is set to Vanilla, FileSystemDomain is set
850 equal to the submit machine's FileSystemDomain.
851 View the requirements of a job which has already been submit‐
853 ted (along with everything else about the job ClassAd) with
854 the command condor_q -l; see the command reference for
855 /man-pages/condor_q. Also, see the HTCondor Users Manual for
856 complete information on the syntax and available attributes
857 that can be used in the ClassAd expression.
858 FILE TRANSFER COMMANDS
860 dont_encrypt_input_files = < file1,file2,file... >
864 A comma and/or space separated list of input files that are
865 not to be network encrypted when transferred with the file
866 transfer mechanism. Specification of files in this manner
867 overrides configuration that would use encryption. Each input
868 file must also be in the list given by transfer_input_files
869 . When a path to an input file or directory is specified,
870 this specifies the path to the file on the submit side. A
871 single wild card character (*) may be used in each file name.
872 dont_encrypt_output_files = < file1,file2,file... >
876 A comma and/or space separated list of output files that are
877 not to be network encrypted when transferred back with the
878 file transfer mechanism. Specification of files in this man‐
879 ner overrides configuration that would use encryption. The
880 output file(s) must also either be in the list given by
881 transfer_output_files or be discovered and to be transferred
882 back with the file transfer mechanism. When a path to an out‐
883 put file or directory is specified, this specifies the path
884 to the file on the execute side. A single wild card character
885 (*) may be used in each file name.
886 encrypt_execute_directory = <True | False>
889 Defaults to False. If set to True, HTCondor will encrypt the
890 contents of the remote scratch directory of the machine where
891 the job is executed. This encryption is transparent to the
892 job itself, but ensures that files left behind on the local
893 disk of the execute machine, perhaps due to a system crash,
894 will remain private. In addition, condor_submit will append
895 to the job's requirements expression
896 && (TARGET.HasEncryptExecuteDirectory)
898 to ensure the job is matched to a machine that is capable of
900 encrypting the contents of the execute directory. This sup‐
901 port is limited to Windows platforms that use the NTFS file
902 system and Linux platforms with the ecryptfs-utils package
903 installed.
904 encrypt_input_files = < file1,file2,file... >
908 A comma and/or space separated list of input files that are
909 to be network encrypted when transferred with the file trans‐
910 fer mechanism. Specification of files in this manner over‐
911 rides configuration that would not use encryption. Each input
912 file must also be in the list given by transfer_input_files
913 . When a path to an input file or directory is specified,
914 this specifies the path to the file on the submit side. A
915 single wild card character (*) may be used in each file name.
916 The method of encryption utilized will be as agreed upon in
917 security negotiation; if that negotiation failed, then the
918 file transfer mechanism must also fail for files to be net‐
919 work encrypted.
920 encrypt_output_files = < file1,file2,file... >
924 A comma and/or space separated list of output files that are
925 to be network encrypted when transferred back with the file
926 transfer mechanism. Specification of files in this manner
927 overrides configuration that would not use encryption. The
928 output file(s) must also either be in the list given by
929 transfer_output_files or be discovered and to be transferred
930 back with the file transfer mechanism. When a path to an out‐
931 put file or directory is specified, this specifies the path
932 to the file on the execute side. A single wild card character
933 (*) may be used in each file name. The method of encryption
934 utilized will be as agreed upon in security negotiation; if
935 that negotiation failed, then the file transfer mechanism
936 must also fail for files to be network encrypted.
937 max_transfer_input_mb = <ClassAd Integer Expression>
940 This integer expression specifies the maximum allowed total
941 size in MiB of the input files that are transferred for a
942 job. This expression does not apply to grid universe, stan‐
943 dard universe, or files transferred via file transfer
944 plug-ins. The expression may refer to attributes of the job.
945 The special value -1 indicates no limit. If not defined, the
946 value set by configuration variable MAX_TRANSFER_INPUT_MB
947 is used. If the observed size of all input files at submit
948 time is larger than the limit, the job will be immediately
949 placed on hold with a HoldReasonCode value of 32. If the job
950 passes this initial test, but the size of the input files in‐
951 creases or the limit decreases so that the limit is violated,
952 the job will be placed on hold at the time when the file
953 transfer is attempted.
954 max_transfer_output_mb = <ClassAd Integer Expression>
957 This integer expression specifies the maximum allowed total
958 size in MiB of the output files that are transferred for a
959 job. This expression does not apply to grid universe, stan‐
960 dard universe, or files transferred via file transfer
961 plug-ins. The expression may refer to attributes of the job.
962 The special value -1 indicates no limit. If not set, the
963 value set by configuration variable MAX_TRANSFER_OUTPUT_MB
964 is used. If the total size of the job's output files to be
965 transferred is larger than the limit, the job will be placed
966 on hold with a HoldReasonCode value of 33. The output will be
967 transferred up to the point when the limit is hit, so some
968 files may be fully transferred, some partially, and some not
969 at all.
970 output_destination = <destination-URL>
974 When present, defines a URL that specifies both a plug-in and
975 a destination for the transfer of the entire output sandbox
976 or a subset of output files as specified by the submit com‐
977 mand transfer_output_files . The plug-in does the transfer
978 of files, and no files are sent back to the submit machine.
979 The HTCondor Administrator's manual has full details.
980 should_transfer_files = <YES | NO | IF_NEEDED >
983 The should_transfer_files setting is used to define if HTCon‐
984 dor should transfer files to and from the remote machine
985 where the job runs. The file transfer mechanism is used to
986 run jobs which are not in the standard universe (and can
987 therefore use remote system calls for file access) on ma‐
988 chines which do not have a shared file system with the submit
989 machine. should_transfer_files equal to YES will cause HT‐
990 Condor to always transfer files for the job. NO disables HT‐
991 Condor's file transfer mechanism. IF_NEEDED will not transfer
992 files for the job if it is matched with a resource in the
993 same FileSystemDomain as the submit machine (and therefore,
994 on a machine with the same shared file system). If the job is
995 matched with a remote resource in a different FileSystemDo‐
996 main, HTCondor will transfer the necessary files.
997 For more information about this and other settings related to
999 transferring files, see the HTCondor User's manual section on
1000 the file transfer mechanism.
1001 Note that should_transfer_files is not supported for jobs
1003 submitted to the grid universe.
1004 skip_filechecks = <True | False>
1007 When True, file permission checks for the submitted job are
1008 disabled. When False, file permissions are checked; this is
1009 the behavior when this command is not present in the submit
1010 description file. File permissions are checked for read per‐
1011 missions on all input files, such as those defined by com‐
1012 mands input and transfer_input_files , and for write per‐
1013 mission to output files, such as a log file defined by log
1014 and output files defined with output or transfer_out‐
1015 put_files .
1016 stream_error = <True | False>
1019 If True, then stderr is streamed back to the machine from
1020 which the job was submitted. If False, stderr is stored lo‐
1021 cally and transferred back when the job completes. This com‐
1022 mand is ignored if the job ClassAd attribute TransferErr is
1023 False. The default value is False. This command must be used
1024 in conjunction with error , otherwise stderr will sent to
1025 /dev/null on Unix machines and ignored on Windows machines.
1026 stream_input = <True | False>
1029 If True, then stdin is streamed from the machine on which the
1030 job was submitted. The default value is False. The command is
1031 only relevant for jobs submitted to the vanilla or java uni‐
1032 verses, and it is ignored by the grid universe. This command
1033 must be used in conjunction with input , otherwise stdin
1034 will be /dev/null on Unix machines and ignored on Windows ma‐
1035 chines.
1036 stream_output = <True | False>
1038 If True, then stdout is streamed back to the machine from
1039 which the job was submitted. If False, stdout is stored lo‐
1040 cally and transferred back when the job completes. This com‐
1041 mand is ignored if the job ClassAd attribute TransferOut is
1042 False. The default value is False. This command must be used
1043 in conjunction with output , otherwise stdout will sent to
1044 /dev/null on Unix machines and ignored on Windows machines.
1045 transfer_executable = <True | False>
1048 This command is applicable to jobs submitted to the grid and
1049 vanilla universes. If transfer_executable is set to False,
1050 then HTCondor looks for the executable on the remote machine,
1051 and does not transfer the executable over. This is useful for
1052 an already pre-staged executable; HTCondor behaves more like
1053 rsh. The default value is True.
1054 transfer_input_files = < file1,file2,file... >
1057 A comma-delimited list of all the files and directories to be
1058 transferred into the working directory for the job, before
1059 the job is started. By default, the file specified in the ex‐
1060 ecutable command and any file specified in the input com‐
1061 mand (for example, stdin) are transferred.
1062 When a path to an input file or directory is specified, this
1064 specifies the path to the file on the submit side. The file
1065 is placed in the job's temporary scratch directory on the ex‐
1066 ecute side, and it is named using the base name of the origi‐
1067 nal path. For example, /path/to/input_file becomes input_file
1068 in the job's scratch directory.
1069 A directory may be specified by appending the forward slash
1071 character (/) as a trailing path separator. This syntax is
1072 used for both Windows and Linux submit hosts. A directory ex‐
1073 ample using a trailing path separator is input_data/. When a
1074 directory is specified with the trailing path separator, the
1075 contents of the directory are transferred, but the directory
1076 itself is not transferred. It is as if each of the items
1077 within the directory were listed in the transfer list. When
1078 there is no trailing path separator, the directory is trans‐
1079 ferred, its contents are transferred, and these contents are
1080 placed inside the transferred directory.
1081 For grid universe jobs other than HTCondor-C, the transfer of
1083 directories is not currently supported.
1084 Symbolic links to files are transferred as the files they
1086 point to. Transfer of symbolic links to directories is not
1087 currently supported.
1088 For vanilla and vm universe jobs only, a file may be speci‐
1090 fied by giving a URL, instead of a file name. The implementa‐
1091 tion for URL transfers requires both configuration and avail‐
1092 able plug-in.
1093 transfer_output_files = < file1,file2,file... >
1096 This command forms an explicit list of output files and di‐
1097 rectories to be transferred back from the temporary working
1098 directory on the execute machine to the submit machine. If
1099 there are multiple files, they must be delimited with commas.
1100 Setting transfer_output_files to the empty string ("") means
1101 that no files are to be transferred.
1102 For HTCondor-C jobs and all other non-grid universe jobs, if
1104 transfer_output_files is not specified, HTCondor will auto‐
1105 matically transfer back all files in the job's temporary
1106 working directory which have been modified or created by the
1107 job. Subdirectories are not scanned for output, so if output
1108 from subdirectories is desired, the output list must be ex‐
1109 plicitly specified. For grid universe jobs other than HTCon‐
1110 dor-C, desired output files must also be explicitly listed.
1111 Another reason to explicitly list output files is for a job
1112 that creates many files, and the user wants only a subset
1113 transferred back.
1114 For grid universe jobs other than with grid type condor, to
1116 have files other than standard output and standard error
1117 transferred from the execute machine back to the submit ma‐
1118 chine, do use transfer_output_files, listing all files to be
1119 transferred. These files are found on the execute machine in
1120 the working directory of the job.
1121 When a path to an output file or directory is specified, it
1123 specifies the path to the file on the execute side. As a des‐
1124 tination on the submit side, the file is placed in the job's
1125 initial working directory, and it is named using the base
1126 name of the original path. For example, path/to/output_file
1127 becomes output_file in the job's initial working directory.
1128 The name and path of the file that is written on the submit
1129 side may be modified by using transfer_output_remaps . Note
1130 that this remap function only works with files but not with
1131 directories.
1132 A directory may be specified using a trailing path separator.
1134 An example of a trailing path separator is the slash charac‐
1135 ter on Unix platforms; a directory example using a trailing
1136 path separator is input_data/. When a directory is specified
1137 with a trailing path separator, the contents of the directory
1138 are transferred, but the directory itself is not transferred.
1139 It is as if each of the items within the directory were
1140 listed in the transfer list. When there is no trailing path
1141 separator, the directory is transferred, its contents are
1142 transferred, and these contents are placed inside the trans‐
1143 ferred directory.
1144 For grid universe jobs other than HTCondor-C, the transfer of
1146 directories is not currently supported.
1147 Symbolic links to files are transferred as the files they
1149 point to. Transfer of symbolic links to directories is not
1150 currently supported.
1151 transfer_output_remaps = < name = newname ; name2 = newname2 ... >
1153 This specifies the name (and optionally path) to use when
1154 downloading output files from the completed job. Normally,
1155 output files are transferred back to the initial working di‐
1156 rectory with the same name they had in the execution direc‐
1157 tory. This gives you the option to save them with a different
1158 path or name. If you specify a relative path, the final path
1159 will be relative to the job's initial working directory.
1160 name describes an output file name produced by your job, and
1162 newname describes the file name it should be downloaded to.
1163 Multiple remaps can be specified by separating each with a
1164 semicolon. If you wish to remap file names that contain
1165 equals signs or semicolons, these special characters may be
1166 escaped with a backslash. You cannot specify directories to
1167 be remapped.
1168 Note that whether an output file is transferred is controlled
1170 by transfer_output_files. Listing a file in transfer_out‐
1171 put_remaps is not sufficient to cause it to be transferred.
1172 when_to_transfer_output = < ON_EXIT | ON_EXIT_OR_EVICT >
1175 Setting when_to_transfer_output equal to ON_EXIT will cause
1176 HTCondor to transfer the job's output files back to the sub‐
1177 mitting machine only when the job completes (exits on its
1178 own).
1179 The ON_EXIT_OR_EVICT option is intended for fault tolerant
1181 jobs which periodically save their own state and can restart
1182 where they left off. In this case, files are spooled to the
1183 submit machine any time the job leaves a remote site, either
1184 because it exited on its own, or was evicted by the HTCondor
1185 system for any reason prior to job completion. The files
1186 spooled back are placed in a directory defined by the value
1187 of the SPOOL configuration variable. Any output files trans‐
1188 ferred back to the submit machine are automatically sent back
1189 out again as input files if the job restarts.
1190 POLICY COMMANDS
1192 max_retries = <integer>
1194 The maximum number of retries allowed for this job (must be
1195 non-negative). If the job fails (does not exit with the suc‐
1196 cess_exit_code exit code) it will be retried up to max_re‐
1197 tries times (unless retries are ceased because of the
1198 retry_until command). If max_retries is not defined, and ei‐
1199 ther retry_until or success_exit_code is, the value of DE‐
1200 FAULT_JOB_MAX_RETRIES will be used for the maximum number of
1201 retries.
1202 The combination of the max_retries, retry_until, and suc‐
1204 cess_exit_code commands causes an appropriate OnExitRemove
1205 expression to be automatically generated. If retry command(s)
1206 and on_exit_remove are both defined, the OnExitRemove expres‐
1207 sion will be generated by OR'ing the expression specified in
1208 OnExitRemove and the expression generated by the retry com‐
1209 mands.
1210 retry_until <Integer | ClassAd Boolean Expression>
1213 An integer value or boolean expression that prevents further
1214 retries from taking place, even if max_retries have not been
1215 exhausted. If retry_until is an integer, the job exiting
1216 with that exit code will cause retries to cease. If retry_un‐
1217 til is a ClassAd expression, the expression evaluating to
1218 True will cause retries to cease.
1219 success_exit_code = <integer>
1221 The exit code that is considered successful for this job. De‐
1222 faults to 0 if not defined.
1223 Note: non-zero values of success_exit_code should generally
1225 not be used for DAG node jobs. At the present time, con‐
1226 dor_dagman does not take into account the value of suc‐
1227 cess_exit_code. This means that, if success_exit_code is set
1228 to a non-zero value, condor_dagman will consider the job
1229 failed when it actually succeeds. For single-proc DAG node
1230 jobs, this can be overcome by using a POST script that takes
1231 into account the value of success_exit_code (although this is
1232 not recommended). For multi-proc DAG node jobs, there is cur‐
1233 rently no way to overcome this limitation.
1234 hold = <True | False>
1237 If hold is set to True, then the submitted job will be placed
1238 into the Hold state. Jobs in the Hold state will not run un‐
1239 til released by condor_release. Defaults to False.
1240 keep_claim_idle = <integer>
1243 An integer number of seconds that a job requests the con‐
1244 dor_schedd to wait before releasing its claim after the job
1245 exits or after the job is removed.
1246 The process by which the condor_schedd claims a condor_startd
1248 is somewhat time-consuming. To amortize this cost, the con‐
1249 dor_schedd tries to reuse claims to run subsequent jobs, af‐
1250 ter a job using a claim is done. However, it can only do this
1251 if there is an idle job in the queue at the moment the previ‐
1252 ous job completes. Sometimes, and especially for the node
1253 jobs when using DAGMan, there is a subsequent job about to be
1254 submitted, but it has not yet arrived in the queue when the
1255 previous job completes. As a result, the condor_schedd re‐
1256 leases the claim, and the next job must wait an entire nego‐
1257 tiation cycle to start. When this submit command is defined
1258 with a non-negative integer, when the job exits, the con‐
1259 dor_schedd tries as usual to reuse the claim. If it cannot,
1260 instead of releasing the claim, the condor_schedd keeps the
1261 claim until either the number of seconds given as a parame‐
1262 ter, or a new job which matches that claim arrives, whichever
1263 comes first. The condor_startd in question will remain in the
1264 Claimed/Idle state, and the original job will be "charged"
1265 (in terms of priority) for the time in this state.
1266 leave_in_queue = <ClassAd Boolean Expression>
1269 When the ClassAd Expression evaluates to True, the job is not
1270 removed from the queue upon completion. This allows the user
1271 of a remotely spooled job to retrieve output files in cases
1272 where HTCondor would have removed them as part of the cleanup
1273 associated with completion. The job will only exit the queue
1274 once it has been marked for removal (via condor_rm, for exam‐
1275 ple) and the leave_in_queue expression has become False.
1276 leave_in_queue defaults to False.
1277 As an example, if the job is to be removed once the output is
1279 retrieved with condor_transfer_data, then use
1280 leave_in_queue = (JobStatus == 4) && ((StageOutFinish =?= UNDEFINED) ||\
1282 (StageOutFinish == 0))
1283 next_job_start_delay = <ClassAd Boolean Expression>
1287 This expression specifies the number of seconds to delay af‐
1288 ter starting up this job before the next job is started. The
1289 maximum allowed delay is specified by the HTCondor configura‐
1290 tion variable MAX_NEXT_JOB_START_DELAY
1291 , which defaults to 10 minutes. This command does not apply
1292 to scheduler or local universe jobs.
1293 This command has been historically used to implement a form
1295 of job start throttling from the job submitter's perspective.
1296 It was effective for the case of multiple job submission
1297 where the transfer of extremely large input data sets to the
1298 execute machine caused machine performance to suffer. This
1299 command is no longer useful, as throttling should be accom‐
1300 plished through configuration of the condor_schedd daemon.
1301 on_exit_hold = <ClassAd Boolean Expression>
1304 The ClassAd expression is checked when the job exits, and if
1305 True, places the job into the Hold state. If False (the de‐
1306 fault value when not defined), then nothing happens and the
1307 on_exit_remove expression is checked to determine if that
1308 needs to be applied.
1309 For example: Suppose a job is known to run for a minimum of
1311 an hour. If the job exits after less than an hour, the job
1312 should be placed on hold and an e-mail notification sent, in‐
1313 stead of being allowed to leave the queue.
1314 on_exit_hold = (time() - JobStartDate) < (60 * $(MINUTE))
1316 This expression places the job on hold if it exits for any
1318 reason before running for an hour. An e-mail will be sent to
1319 the user explaining that the job was placed on hold because
1320 this expression became True.
1321 periodic_* expressions take precedence over on_exit_* expres‐
1323 sions, and *_hold expressions take precedence over a *_remove
1324 expressions.
1325 Only job ClassAd attributes will be defined for use by this
1327 ClassAd expression. This expression is available for the
1328 vanilla, java, parallel, grid, local and scheduler universes.
1329 It is additionally available, when submitted from a Unix ma‐
1330 chine, for the standard universe.
1331 on_exit_hold_reason = <ClassAd String Expression>
1333 When the job is placed on hold due to the on_exit_hold ex‐
1334 pression becoming True, this expression is evaluated to set
1335 the value of HoldReason in the job ClassAd. If this expres‐
1336 sion is UNDEFINED or produces an empty or invalid string, a
1337 default description is used.
1338 on_exit_hold_subcode = <ClassAd Integer Expression>
1341 When the job is placed on hold due to the on_exit_hold ex‐
1342 pression becoming True, this expression is evaluated to set
1343 the value of HoldReasonSubCode in the job ClassAd. The de‐
1344 fault subcode is 0. The HoldReasonCode will be set to 3,
1345 which indicates that the job went on hold due to a job policy
1346 expression.
1347 on_exit_remove = <ClassAd Boolean Expression>
1350 The ClassAd expression is checked when the job exits, and if
1351 True (the default value when undefined), then it allows the
1352 job to leave the queue normally. If False, then the job is
1353 placed back into the Idle state. If the user job runs under
1354 the vanilla universe, then the job restarts from the begin‐
1355 ning. If the user job runs under the standard universe, then
1356 it continues from where it left off, using the last check‐
1357 point.
1358 For example, suppose a job occasionally segfaults, but
1360 chances are that the job will finish successfully if the job
1361 is run again with the same data. The on_exit_remove expres‐
1362 sion can cause the job to run again with the following com‐
1363 mand. Assume that the signal identifier for the segmentation
1364 fault is 11 on the platform where the job will be running.
1365 on_exit_remove = (ExitBySignal == False) || (ExitSignal != 11)
1367 This expression lets the job leave the queue if the job was
1369 not killed by a signal or if it was killed by a signal other
1370 than 11, representing segmentation fault in this example. So,
1371 if the exited due to signal 11, it will stay in the job
1372 queue. In any other case of the job exiting, the job will
1373 leave the queue as it normally would have done.
1374 As another example, if the job should only leave the queue if
1376 it exited on its own with status 0, this on_exit_remove ex‐
1377 pression works well:
1378 on_exit_remove = (ExitBySignal == False) && (ExitCode == 0)
1380 If the job was killed by a signal or exited with a non-zero
1382 exit status, HTCondor would leave the job in the queue to run
1383 again.
1384 periodic_* expressions take precedence over on_exit_* expres‐
1386 sions, and *_hold expressions take precedence over a *_remove
1387 expressions.
1388 Only job ClassAd attributes will be defined for use by this
1390 ClassAd expression.
1391 periodic_hold = <ClassAd Boolean Expression>
1393 This expression is checked periodically when the job is not
1394 in the Held state. If it becomes True, the job will be placed
1395 on hold. If unspecified, the default value is False.
1396 periodic_* expressions take precedence over on_exit_* expres‐
1398 sions, and *_hold expressions take precedence over a *_remove
1399 expressions.
1400 Only job ClassAd attributes will be defined for use by this
1402 ClassAd expression. Note that, by default, this expression is
1403 only checked once every 60 seconds. The period of these eval‐
1404 uations can be adjusted by setting the PERIODIC_EXPR_INTER‐
1405 VAL, MAX_PERIODIC_EXPR_INTERVAL, and PERIODIC_EXPR_TIMESLICE
1406 configuration macros.
1407 periodic_hold_reason = <ClassAd String Expression>
1410 When the job is placed on hold due to the periodic_hold ex‐
1411 pression becoming True, this expression is evaluated to set
1412 the value of HoldReason in the job ClassAd. If this expres‐
1413 sion is UNDEFINED or produces an empty or invalid string, a
1414 default description is used.
1415 periodic_hold_subcode = <ClassAd Integer Expression>
1418 When the job is placed on hold due to the periodic_hold ex‐
1419 pression becoming true, this expression is evaluated to set
1420 the value of HoldReasonSubCode in the job ClassAd. The de‐
1421 fault subcode is 0. The HoldReasonCode will be set to 3,
1422 which indicates that the job went on hold due to a job policy
1423 expression.
1424 periodic_release = <ClassAd Boolean Expression>
1427 This expression is checked periodically when the job is in
1428 the Held state. If the expression becomes True, the job will
1429 be released.
1430 Only job ClassAd attributes will be defined for use by this
1432 ClassAd expression. Note that, by default, this expression is
1433 only checked once every 60 seconds. The period of these eval‐
1434 uations can be adjusted by setting the PERIODIC_EXPR_INTER‐
1435 VAL, MAX_PERIODIC_EXPR_INTERVAL, and PERIODIC_EXPR_TIMESLICE
1436 configuration macros.
1437 periodic_remove = <ClassAd Boolean Expression>
1440 This expression is checked periodically. If it becomes True,
1441 the job is removed from the queue. If unspecified, the de‐
1442 fault value is False.
1443 See the Examples section of this manual page for an example
1445 of a periodic_remove expression.
1446 periodic_* expressions take precedence over on_exit_* expres‐
1448 sions, and *_hold expressions take precedence over a *_remove
1449 expressions. So, the periodic_remove expression takes prece‐
1450 dent over the on_exit_remove expression, if the two describe
1451 conflicting actions.
1452 Only job ClassAd attributes will be defined for use by this
1454 ClassAd expression. Note that, by default, this expression is
1455 only checked once every 60 seconds. The period of these eval‐
1456 uations can be adjusted by setting the PERIODIC_EXPR_INTER‐
1457 VAL, MAX_PERIODIC_EXPR_INTERVAL, and PERIODIC_EXPR_TIMESLICE
1458 configuration macros.
1459 COMMANDS SPECIFIC TO THE STANDARD UNIVERSE
1461 allow_startup_script = <True | False>
1464 If True, a standard universe job will execute a script in‐
1465 stead of submitting the job, and the consistency check to see
1466 if the executable has been linked using condor_compile is
1467 omitted. The executable command within the submit descrip‐
1468 tion file specifies the name of the script. The script is
1469 used to do preprocessing before the job is submitted. The
1470 shell script ends with an exec of the job executable, such
1471 that the process id of the executable is the same as that of
1472 the shell script. Here is an example script that gets a copy
1473 of a machine-specific executable before the exec.
1474 #! /bin/sh
1476 # get the host name of the machine
1478 $host=`uname -n`
1479 # grab a standard universe executable designed specifically
1481 # for this host
1482 scp elsewhere@cs.wisc.edu:${host} executable
1483 # The PID MUST stay the same, so exec the new standard universe process.
1485 exec executable ${1+"$@"}
1486 If this command is not present (defined), then the value de‐
1488 faults to false.
1489 append_files = file1, file2, ...
1491 If your job attempts to access a file mentioned in this list,
1492 HTCondor will force all writes to that file to be appended to
1493 the end. Furthermore, condor_submit will not truncate it.
1494 This list uses the same syntax as compress_files, shown
1495 above.
1496 This option may yield some surprising results. If several
1498 jobs attempt to write to the same file, their output may be
1499 intermixed. If a job is evicted from one or more machines
1500 during the course of its lifetime, such an output file might
1501 contain several copies of the results. This option should be
1502 only be used when you wish a certain file to be treated as a
1503 running log instead of a precise result.
1504 This option only applies to standard-universe jobs.
1506 buffer_files = < name = (size,block-size) ; name2 =
1511 (size,block-size) ... >; buffer_size = <bytes-in-buffer>; buf‐
1512 fer_block_size = <bytes-in-block>
1513 HTCondor keeps a buffer of recently-used data for each file a
1514 job accesses. This buffer is used both to cache commonly-used
1515 data and to consolidate small reads and writes into larger
1516 operations that get better throughput. The default settings
1517 should produce reasonable results for most programs.
1518 These options only apply to standard-universe jobs.
1520 If needed, you may set the buffer controls individually for
1522 each file using the buffer_files option. For example, to set
1523 the buffer size to 1 MiB and the block size to 256 KiB for
1524 the file input.data, use this command:
1525 buffer_files = "input.data=(1000000,256000)"
1527 Alternatively, you may use these two options to set the de‐
1529 fault sizes for all files used by your job:
1530 buffer_size = 1000000
1532 buffer_block_size = 256000
1533 If you do not set these, HTCondor will use the values given
1535 by these two configuration file macros:
1536 DEFAULT_IO_BUFFER_SIZE = 1000000
1538 DEFAULT_IO_BUFFER_BLOCK_SIZE = 256000
1539 Finally, if no other settings are present, HTCondor will use
1541 a buffer of 512 KiB and a block size of 32 KiB.
1542 compress_files = file1, file2, ...
1545 If your job attempts to access any of the files mentioned in
1546 this list, HTCondor will automatically compress them (if
1547 writing) or decompress them (if reading). The compress format
1548 is the same as used by GNU gzip.
1549 The files given in this list may be simple file names or com‐
1551 plete paths and may include * as a wild card. For example,
1552 this list causes the file /tmp/data.gz, any file named
1553 event.gz, and any file ending in .gzip to be automatically
1554 compressed or decompressed as needed:
1555 compress_files = /tmp/data.gz, event.gz, *.gzip
1557 Due to the nature of the compression format, compressed files
1559 must only be accessed sequentially. Random access reading is
1560 allowed but is very slow, while random access writing is sim‐
1561 ply not possible. This restriction may be avoided by using
1562 both compress_files and fetch_files at the same time. When
1563 this is done, a file is kept in the decompressed state at the
1564 execution machine, but is compressed for transfer to its
1565 original location.
1566 This option only applies to standard universe jobs.
1568 fetch_files = file1, file2, ...
1571 If your job attempts to access a file mentioned in this list,
1572 HTCondor will automatically copy the whole file to the exe‐
1573 cuting machine, where it can be accessed quickly. When your
1574 job closes the file, it will be copied back to its original
1575 location. This list uses the same syntax as compress_files,
1576 shown above.
1577 This option only applies to standard universe jobs.
1579 file_remaps = < name = newname ; name2 = newname2 ... >
1582 Directs HTCondor to use a new file name in place of an old
1583 one. name describes a file name that your job may attempt to
1584 open, and newname describes the file name it should be re‐
1585 placed with. newname may include an optional leading access
1586 specifier, local: or remote:. If left unspecified, the de‐
1587 fault access specifier is remote:. Multiple remaps can be
1588 specified by separating each with a semicolon.
1589 This option only applies to standard universe jobs.
1591 If you wish to remap file names that contain equals signs or
1593 semicolons, these special characters may be escaped with a
1594 backslash.
1595 Example One:
1597 Suppose that your job reads a file named
1598 dataset.1. To instruct HTCondor to force your job
1599 to read other.dataset instead, add this to the
1600 submit file:
1601 file_remaps = "dataset.1=other.dataset"
1603 Example Two:
1605 Suppose that your run many jobs which all read in
1606 the same large file, called very.big. If this file
1607 can be found in the same place on a local disk in
1608 every machine in the pool, (say /bigdisk/bigfile,)
1609 you can instruct HTCondor of this fact by remap‐
1610 ping very.big to /bigdisk/bigfile and specifying
1611 that the file is to be read locally, which will be
1612 much faster than reading over the network.
1613 file_remaps = "very.big = local:/bigdisk/bigfile"
1615 Example Three:
1617 Several remaps can be applied at once by separat‐
1618 ing each with a semicolon.
1619 file_remaps = "very.big = local:/bigdisk/bigfile ; dataset.1 = other.dataset"
1621 local_files = file1, file2, ...
1625 If your job attempts to access a file mentioned in this list,
1626 HTCondor will cause it to be read or written at the execution
1627 machine. This is most useful for temporary files not used for
1628 input or output. This list uses the same syntax as com‐
1629 press_files, shown above.
1630 local_files = /tmp/*
1632 This option only applies to standard universe jobs.
1634 want_remote_io = <True | False>
1637 This option controls how a file is opened and manipulated in
1638 a standard universe job. If this option is true, which is the
1639 default, then the condor_shadow makes all decisions about how
1640 each and every file should be opened by the executing job.
1641 This entails a network round trip (or more) from the job to
1642 the condor_shadow and back again for every single open() in
1643 addition to other needed information about the file. If set
1644 to false, then when the job queries the condor_shadow for the
1645 first time about how to open a file, the condor_shadow will
1646 inform the job to automatically perform all of its file ma‐
1647 nipulation on the local file system on the execute machine
1648 and any file remapping will be ignored. This means that there
1649 must be a shared file system (such as NFS or AFS) between the
1650 execute machine and the submit machine and that ALL paths
1651 that the job could open on the execute machine must be valid.
1652 The ability of the standard universe job to checkpoint, pos‐
1653 sibly to a checkpoint server, is not affected by this attri‐
1654 bute. However, when the job resumes it will be expecting the
1655 same file system conditions that were present when the job
1656 checkpointed.
1657 COMMANDS FOR THE GRID
1659 azure_admin_key = <pathname>
1661 For grid type azure jobs, specifies the path and file name of
1662 a file that contains an SSH public key. This key can be used
1663 to log into the administrator account of the instance via
1664 SSH.
1665 azure_admin_username = <account name>
1668 For grid type azure jobs, specifies the name of an adminis‐
1669 trator account to be created in the instance. This account
1670 can be logged into via SSH.
1671 azure_auth_file = <pathname>
1673 For grid type azure jobs, specifies a path and file name of
1674 the authorization file that grants permission for HTCondor to
1675 use the Azure account. If it's not defined, then HTCondor
1676 will attempt to use the default credentials of the Azure CLI
1677 tools.
1678 azure_image = <image id>
1681 For grid type azure jobs, identifies the disk image to be
1682 used for the boot disk of the instance. This image must al‐
1683 ready be registered within Azure.
1684 azure_location = <image id>
1687 For grid type azure jobs, identifies the location within
1688 Azure where the instance should be run. As an example, one
1689 current location is centralus.
1690 azure_size = <machine type>
1693 For grid type azure jobs, the hardware configuration that the
1694 virtual machine instance is to run on.
1695 batch_queue = <queuename>
1698 Used for pbs, lsf, and sge grid universe jobs. Specifies the
1699 name of the PBS/LSF/SGE job queue into which the job should
1700 be submitted. If not specified, the default queue is used.
1701 boinc_authenticator_file = <pathname>
1704 For grid type boinc jobs, specifies a path and file name of
1705 the authorization file that grants permission for HTCondor to
1706 use the BOINC service. There is no default value when not
1707 specified.
1708 cream_attributes = <name=value;...;name=value>
1711 Provides a list of attribute/value pairs to be set in a CREAM
1712 job description of a grid universe job destined for the CREAM
1713 grid system. The pairs are separated by semicolons, and writ‐
1714 ten in New ClassAd syntax.
1715 delegate_job_GSI_credentials_lifetime = <seconds>
1718 Specifies the maximum number of seconds for which delegated
1719 proxies should be valid. The default behavior when this com‐
1720 mand is not specified is determined by the configuration
1721 variable DELEGATE_JOB_GSI_CREDENTIALS_LIFETIME
1722 , which defaults to one day. A value of 0 indicates that the
1723 delegated proxy should be valid for as long as allowed by the
1724 credential used to create the proxy. This setting currently
1725 only applies to proxies delegated for non-grid jobs and for
1726 HTCondor-C jobs. It does not currently apply to globus grid
1727 jobs, which always behave as though this setting were 0. This
1728 variable has no effect if the configuration variable DELE‐
1729 GATE_JOB_GSI_CREDENTIALS
1730 is False, because in that case the job proxy is copied
1731 rather than delegated.
1732 ec2_access_key_id = <pathname>
1735 For grid type ec2 jobs, identifies the file containing the
1736 access key.
1737 ec2_ami_id = <EC2 xMI ID>
1739 For grid type ec2 jobs, identifies the machine image. Ser‐
1740 vices compatible with the EC2 Query API may refer to these
1741 with abbreviations other than AMI, for example EMI is valid
1742 for Eucalyptus.
1743 ec2_availability_zone = <zone name>
1745 For grid type ec2 jobs, specifies the Availability Zone that
1746 the instance should be run in. This command is optional, un‐
1747 less ec2_ebs_volumes is set. As an example, one current zone
1748 is us-east-1b.
1749 ec2_block_device_mapping = <block-device>:<kernel-device>,<block-de‐
1752 vice>:<kernel-device>, ...
1753 For grid type ec2 jobs, specifies the block device to kernel
1754 device mapping. This command is optional.
1755 ec2_ebs_volumes = <ebs name>:<device name>,<ebs name>:<device
1758 name>,...
1759 For grid type ec2 jobs, optionally specifies a list of Elas‐
1760 tic Block Store (EBS) volumes to be made available to the in‐
1761 stance and the device names they should have in the instance.
1762 ec2_elastic_ip = <elastic IP address>
1765 For grid type ec2 jobs, and optional specification of an
1766 Elastic IP address that should be assigned to this instance.
1767 ec2_iam_profile_arn = <IAM profile ARN>
1770 For grid type ec2 jobs, an Amazon Resource Name (ARN) identi‐
1771 fying which Identity and Access Management (IAM) (instance)
1772 profile to associate with the instance.
1773 ec2_iam_profile_name= <IAM profile name>
1776 For grid type ec2 jobs, a name identifying which Identity and
1777 Access Management (IAM) (instance) profile to associate with
1778 the instance.
1779 ec2_instance_type = <instance type>
1781 For grid type ec2 jobs, identifies the instance type. Differ‐
1782 ent services may offer different instance types, so no de‐
1783 fault value is set.
1784 ec2_keypair = <ssh key-pair name>
1786 For grid type ec2 jobs, specifies the name of an SSH key-pair
1787 that is already registered with the EC2 service. The associ‐
1788 ated private key can be used to ssh into the virtual machine
1789 once it is running.
1790 ec2_keypair_file = <pathname>
1792 For grid type ec2 jobs, specifies the complete path and file
1793 name of a file into which HTCondor will write an SSH key for
1794 use with ec2 jobs. The key can be used to ssh into the vir‐
1795 tual machine once it is running. If ec2_keypair is speci‐
1796 fied for a job, ec2_keypair_file is ignored.
1797 ec2_parameter_names = ParameterName1, ParameterName2, ...
1799 For grid type ec2 jobs, a space or comma separated list of
1800 the names of additional parameters to pass when instantiating
1801 an instance.
1802 ec2_parameter_<name> = <value>
1804 For grid type ec2 jobs, specifies the value for the corre‐
1805 spondingly named (instance instantiation) parameter. <name>
1806 is the parameter name specified in the submit command ec2_pa‐
1807 rameter_names , but with any periods replaced by under‐
1808 scores.
1809 ec2_secret_access_key = <pathname>
1812 For grid type ec2 jobs, specifies the path and file name con‐
1813 taining the secret access key.
1814 ec2_security_groups = group1, group2, ...
1817 For grid type ec2 jobs, defines the list of EC2 security
1818 groups which should be associated with the job.
1819 ec2_security_ids = id1, id2, ...
1822 For grid type ec2 jobs, defines the list of EC2 security
1823 group IDs which should be associated with the job.
1824 ec2_spot_price = <bid>
1827 For grid type ec2 jobs, specifies the spot instance bid,
1828 which is the most that the job submitter is willing to pay
1829 per hour to run this job.
1830 ec2_tag_names = <name0,name1,name...>
1832 For grid type ec2 jobs, specifies the case of tag names that
1833 will be associated with the running instance. This is only
1834 necessary if a tag name case matters. By default the list
1835 will be automatically generated.
1836 ec2_tag_<name> = <value>
1839 For grid type ec2 jobs, specifies a tag to be associated with
1840 the running instance. The tag name will be lower-cased, use
1841 ec2_tag_names to change the case.
1842 WantNameTag = <True | False>
1844 For grid type ec2 jobs, a job may request that its 'name' tag
1845 be (not) set by HTCondor. If the job does not otherwise spec‐
1846 ify any tags, not setting its name tag will eliminate a call
1847 by the EC2 GAHP, improving performance.
1848 ec2_user_data = <data>
1851 For grid type ec2 jobs, provides a block of data that can be
1852 accessed by the virtual machine. If both ec2_user_data and
1853 ec2_user_data_file are specified for a job, the two blocks of
1854 data are concatenated, with the data from this ec2_user_data
1855 submit command occurring first.
1856 ec2_user_data_file = <pathname>
1858 For grid type ec2 jobs, specifies a path and file name whose
1859 contents can be accessed by the virtual machine. If both
1860 ec2_user_data and ec2_user_data_file are specified for a job,
1861 the two blocks of data are concatenated, with the data from
1862 that ec2_user_data submit command occurring first.
1863 ec2_vpc_ip = <a.b.c.d>
1865 For grid type ec2 jobs, that are part of a Virtual Private
1866 Cloud (VPC), an optional specification of the IP address that
1867 this instance should have within the VPC.
1868 ec2_vpc_subnet = <subnet specification string>
1871 For grid type ec2 jobs, an optional specification of the Vir‐
1872 tual Private Cloud (VPC) that this instance should be a part
1873 of.
1874 gce_account = <account name>
1877 For grid type gce jobs, specifies the Google cloud services
1878 account to use. If this submit command isn't specified, then
1879 a random account from the authorization file given by
1880 gce_auth_file will be used.
1881 gce_auth_file = <pathname>
1883 For grid type gce jobs, specifies a path and file name of the
1884 authorization file that grants permission for HTCondor to use
1885 the Google account. If this command is not specified, then
1886 the default file of the Google command-line tools will be
1887 used.
1888 gce_image = <image id>
1891 For grid type gce jobs, the identifier of the virtual machine
1892 image representing the HTCondor job to be run. This virtual
1893 machine image must already be register with GCE and reside in
1894 Google's Cloud Storage service.
1895 gce_json_file = <pathname>
1897 For grid type gce jobs, specifies the path and file name of a
1898 file that contains JSON elements that should be added to the
1899 instance description submitted to the GCE service.
1900 gce_machine_type = <machine type>
1903 For grid type gce jobs, the long form of the URL that de‐
1904 scribes the machine configuration that the virtual machine
1905 instance is to run on.
1906 gce_metadata = <name=value,...,name=value>
1908 For grid type gce jobs, a comma separated list of name and
1909 value pairs that define metadata for a virtual machine in‐
1910 stance that is an HTCondor job.
1911 gce_metadata_file = <pathname>
1913 For grid type gce jobs, specifies a path and file name of the
1914 file that contains metadata for a virtual machine instance
1915 that is an HTCondor job. Within the file, each name and value
1916 pair is on its own line; so, the pairs are separated by the
1917 newline character.
1918 gce_preemptible = <True | False>
1921 For grid type gce jobs, specifies whether the virtual machine
1922 instance should be preemptible. The default is for the in‐
1923 stance to not be preemptible.
1924 globus_rematch = <ClassAd Boolean Expression>
1926 This expression is evaluated by the condor_gridmanager when‐
1927 ever:
1928 1. the globus_resubmit expression evaluates to True
1930 2. the condor_gridmanager decides it needs to retry a submis‐
1932 sion (as when a previous submission failed to commit)
1933 If globus_rematch evaluates to True, then before the job is
1935 submitted again to globus, the condor_gridmanager will re‐
1936 quest that the condor_schedd daemon renegotiate with the
1937 matchmaker (the condor_negotiator). The result is this job
1938 will be matched again.
1939 globus_resubmit = <ClassAd Boolean Expression>
1942 The expression is evaluated by the condor_gridmanager each
1943 time the condor_gridmanager gets a job ad to manage. There‐
1944 fore, the expression is evaluated:
1945 1. when a grid universe job is first submitted to HTCondor-G
1947 2. when a grid universe job is released from the hold state
1949 3. when HTCondor-G is restarted (specifically, whenever the
1951 condor_gridmanager is restarted)
1952 If the expression evaluates to True, then any previous sub‐
1954 mission to the grid universe will be forgotten and this job
1955 will be submitted again as a fresh submission to the grid
1956 universe. This may be useful if there is a desire to give up
1957 on a previous submission and try again. Note that this may
1958 result in the same job running more than once. Do not treat
1959 this operation lightly.
1960 globus_rsl = <RSL-string>
1963 Used to provide any additional Globus RSL string attributes
1964 which are not covered by other submit description file com‐
1965 mands or job attributes. Used for grid universe jobs, where
1966 the grid resource has a grid-type-string of gt2.
1967 grid_resource = <grid-type-string> <grid-specific-parameter-list>
1970 For each grid-type-string value, there are further type-spe‐
1971 cific values that must specified. This submit description
1972 file command allows each to be given in a space-separated
1973 list. Allowable grid-type-string values are batch, condor,
1974 cream, ec2, gt2, gt5, lsf, nordugrid, pbs, sge, and unicore.
1975 The HTCondor manual chapter on Grid Computing details the va‐
1976 riety of grid types.
1977 For a grid-type-string of batch, the single parameter is the
1979 name of the local batch system, and will be one of pbs, lsf,
1980 or sge.
1981 For a grid-type-string of condor, the first parameter is the
1983 name of the remote condor_schedd daemon. The second parameter
1984 is the name of the pool to which the remote condor_schedd
1985 daemon belongs.
1986 For a grid-type-string of cream, there are three parameters.
1988 The first parameter is the web services address of the CREAM
1989 server. The second parameter is the name of the batch system
1990 that sits behind the CREAM server. The third parameter iden‐
1991 tifies a site-specific queue within the batch system.
1992 For a grid-type-string of ec2, one additional parameter spec‐
1994 ifies the EC2 URL.
1995 For a grid-type-string of gt2, the single parameter is the
1997 name of the pre-WS GRAM resource to be used.
1998 For a grid-type-string of gt5, the single parameter is the
2000 name of the pre-WS GRAM resource to be used, which is the
2001 same as for the grid-type-string of gt2.
2002 For a grid-type-string of lsf, no additional parameters are
2004 used.
2005 For a grid-type-string of nordugrid, the single parameter is
2007 the name of the NorduGrid resource to be used.
2008 For a grid-type-string of pbs, no additional parameters are
2010 used.
2011 For a grid-type-string of sge, no additional parameters are
2013 used.
2014 For a grid-type-string of unicore, the first parameter is the
2016 name of the Unicore Usite to be used. The second parameter is
2017 the name of the Unicore Vsite to be used.
2018 keystore_alias = <name>
2021 A string to locate the certificate in a Java keystore file,
2022 as used for a unicore job.
2023 keystore_file = <pathname>
2026 The complete path and file name of the Java keystore file
2027 containing the certificate to be used for a unicore job.
2028 keystore_passphrase_file = <pathname>
2031 The complete path and file name to the file containing the
2032 passphrase protecting a Java keystore file containing the
2033 certificate. Relevant for a unicore job.
2034 MyProxyCredentialName = <symbolic name>
2037 The symbolic name that identifies a credential to the MyProxy
2038 server. This symbolic name is set as the credential is ini‐
2039 tially stored on the server (using myproxy-init).
2040 MyProxyHost = <host>:<port>
2043 The Internet address of the host that is the MyProxy server.
2044 The host may be specified by either a host name (as in
2045 head.example.com) or an IP address (of the form 123.456.7.8).
2046 The port number is an integer.
2047 MyProxyNewProxyLifetime = <number-of-minutes>
2050 The new lifetime (in minutes) of the proxy after it is re‐
2051 freshed.
2052 MyProxyPassword = <password>
2055 The password needed to refresh a credential on the MyProxy
2056 server. This password is set when the user initially stores
2057 credentials on the server (using myproxy-init). As an alter‐
2058 native to using MyProxyPassword in the submit description
2059 file, the password may be specified as a command line argu‐
2060 ment to condor_submit with the -password argument.
2061 MyProxyRefreshThreshold = <number-of-seconds>
2063 The time (in seconds) before the expiration of a proxy that
2064 the proxy should be refreshed. For example, if MyProxyRe‐
2065 freshThreshold is set to the value 600, the proxy will be re‐
2066 freshed 10 minutes before it expires.
2067 MyProxyServerDN = <credential subject>
2069 A string that specifies the expected Distinguished Name (cre‐
2070 dential subject, abbreviated DN) of the MyProxy server. It
2071 must be specified when the MyProxy server DN does not follow
2072 the conventional naming scheme of a host credential. This oc‐
2073 curs, for example, when the MyProxy server DN begins with a
2074 user credential.
2075 nordugrid_rsl = <RSL-string>
2078 Used to provide any additional RSL string attributes which
2079 are not covered by regular submit description file parame‐
2080 ters. Used when the universe is grid, and the type of grid
2081 system is nordugrid.
2082 transfer_error = <True | False>
2084 For jobs submitted to the grid universe only. If True, then
2085 the error output (from stderr) from the job is transferred
2086 from the remote machine back to the submit machine. The name
2087 of the file after transfer is given by the error command.
2088 If False, no transfer takes place (from the remote machine to
2089 submit machine), and the name of the file is given by the er‐
2090 ror command. The default value is True.
2091 transfer_input = <True | False>
2094 For jobs submitted to the grid universe only. If True, then
2095 the job input (stdin) is transferred from the machine where
2096 the job was submitted to the remote machine. The name of the
2097 file that is transferred is given by the input command. If
2098 False, then the job's input is taken from a pre-staged file
2099 on the remote machine, and the name of the file is given by
2100 the input command. The default value is True.
2101 For transferring files other than stdin, see transfer_in‐
2103 put_files .
2104 transfer_output = <True | False>
2107 For jobs submitted to the grid universe only. If True, then
2108 the output (from stdout) from the job is transferred from the
2109 remote machine back to the submit machine. The name of the
2110 file after transfer is given by the output command. If
2111 False, no transfer takes place (from the remote machine to
2112 submit machine), and the name of the file is given by the
2113 output command. The default value is True.
2114 For transferring files other than stdout, see transfer_out‐
2116 put_files .
2117 use_x509userproxy = <True | False>
2120 Set this command to True to indicate that the job requires an
2121 X.509 user proxy. If x509userproxy is set, then that file is
2122 used for the proxy. Otherwise, the proxy is looked for in the
2123 standard locations. If x509userproxy is set or if the job is
2124 a grid universe job of grid type gt2, gt5, cream, or nor‐
2125 dugrid, then the value of use_x509userproxy is forced to
2126 True. Defaults to False.
2127 x509userproxy = <full-pathname>
2130 Used to override the default path name for X.509 user cer‐
2131 tificates. The default location for X.509 proxies is the
2132 /tmp directory, which is generally a local file system. Set‐
2133 ting this value would allow HTCondor to access the proxy in a
2134 shared file system (for example, AFS). HTCondor will use the
2135 proxy specified in the submit description file first. If
2136 nothing is specified in the submit description file, it will
2137 use the environment variable X509_USER_PROXY. If that vari‐
2138 able is not present, it will search in the default location.
2139 Note that proxies are only valid for a limited time. Con‐
2140 dor_submit will not submit a job with an expired proxy, it
2141 will return an error. Also, if the configuration parameter
2142 CRED_MIN_TIME_LEFT is set to some number of seconds, and if
2143 the proxy will expire before that many seconds, condor_submit
2144 will also refuse to submit the job. That is, if
2145 CRED_MIN_TIME_LEFT is set to 60, condor_submit will refuse to
2146 submit a job whose proxy will expire 60 seconds from the time
2147 of submission.
2148 x509userproxy is relevant when the universe is vanilla, or
2150 when the universe is grid and the type of grid system is one
2151 of gt2, gt5, condor, cream, or nordugrid. Defining a value
2152 causes the proxy to be delegated to the execute machine.
2153 Further, VOMS attributes defined in the proxy will appear in
2154 the job ClassAd.
2155 COMMANDS FOR PARALLEL, JAVA, and SCHEDULER UNIVERSES
2157 hold_kill_sig = <signal-number>
2160 For the scheduler universe only, signal-number is the sig‐
2161 nal delivered to the job when the job is put on hold with
2162 condor_hold. signal-number may be either the platform-spe‐
2163 cific name or value of the signal. If this command is not
2164 present, the value of kill_sig is used.
2165 jar_files = <file_list>
2168 Specifies a list of additional JAR files to include when us‐
2169 ing the Java universe. JAR files will be transferred along
2170 with the executable and automatically added to the classpath.
2171 java_vm_args = <argument_list>
2174 Specifies a list of additional arguments to the Java VM it‐
2175 self, When HTCondor runs the Java program, these are the ar‐
2176 guments that go before the class name. This can be used to
2177 set VM-specific arguments like stack size, garbage-collector
2178 arguments and initial property values.
2179 machine_count = <max>
2181 For the parallel universe, a single value (max) is required.
2182 It is neither a maximum or minimum, but the number of ma‐
2183 chines to be dedicated toward running the job.
2184 remove_kill_sig = <signal-number>
2187 For the scheduler universe only, signal-number is the sig‐
2188 nal delivered to the job when the job is removed with con‐
2189 dor_rm. signal-number may be either the platform-specific
2190 name or value of the signal. This example shows it both ways
2191 for a Linux signal:
2192 remove_kill_sig = SIGUSR1
2194 remove_kill_sig = 10
2195 If this command is not present, the value of kill_sig is
2197 used.
2198 COMMANDS FOR THE VM UNIVERSE
2200 vm_disk = file1:device1:permission1, file2:device2:permission2:for‐
2202 mat2, ...
2203 A list of comma separated disk files. Each disk file is spec‐
2204 ified by 4 colon separated fields. The first field is the
2205 path and file name of the disk file. The second field speci‐
2206 fies the device. The third field specifies permissions, and
2207 the optional fourth field specifies the image format. If a
2208 disk file will be transferred by HTCondor, then the first
2209 field should just be the simple file name (no path informa‐
2210 tion).
2211 An example that specifies two disk files:
2213 vm_disk = /myxen/diskfile.img:sda1:w,/myxen/swap.img:sda2:w
2215 vm_checkpoint = <True | False>
2219 A boolean value specifying whether or not to take check‐
2220 points. If not specified, the default value is False. In the
2221 current implementation, setting both vm_checkpoint and
2222 vm_networking to True does not yet work in all cases. Net‐
2223 working cannot be used if a vm universe job uses a checkpoint
2224 in order to continue execution after migration to another ma‐
2225 chine.
2226 vm_macaddr = <MACAddr>
2229 Defines that MAC address that the virtual machine's network
2230 interface should have, in the standard format of six groups
2231 of two hexadecimal digits separated by colons.
2232 vm_memory = <MBytes-of-memory>
2235 The amount of memory in MBytes that a vm universe job re‐
2236 quires.
2237 vm_networking = <True | False>
2240 Specifies whether to use networking or not. In the current
2241 implementation, setting both vm_checkpoint and vm_networking
2242 to True does not yet work in all cases. Networking cannot be
2243 used if a vm universe job uses a checkpoint in order to con‐
2244 tinue execution after migration to another machine.
2245 vm_networking_type = <nat | bridge >
2248 When vm_networking is True, this definition augments the
2249 job's requirements to match only machines with the specified
2250 networking. If not specified, then either networking type
2251 matches.
2252 vm_no_output_vm = <True | False>
2255 When True, prevents HTCondor from transferring output files
2256 back to the machine from which the vm universe job was sub‐
2257 mitted. If not specified, the default value is False.
2258 vm_type = <vmware | xen | kvm>
2261 Specifies the underlying virtual machine software that this
2262 job expects.
2263 vmware_dir = <pathname>
2265 The complete path and name of the directory where VMware-spe‐
2266 cific files and applications such as the VMDK (Virtual Ma‐
2267 chine Disk Format) and VMX (Virtual Machine Configuration)
2268 reside. This command is optional; when not specified, all
2269 relevant VMware image files are to be listed using trans‐
2270 fer_input_files .
2271 vmware_should_transfer_files = <True | False>
2274 Specifies whether HTCondor will transfer VMware-specific
2275 files located as specified by vmware_dir to the execute ma‐
2276 chine (True) or rely on access through a shared file system
2277 (False). Omission of this required command (for VMware vm
2278 universe jobs) results in an error message from condor_sub‐
2279 mit, and the job will not be submitted.
2280 vmware_snapshot_disk = <True | False>
2283 When True, causes HTCondor to utilize a VMware snapshot disk
2284 for new or modified files. If not specified, the default
2285 value is True.
2286 xen_initrd = <image-file>
2288 When xen_kernel gives a file name for the kernel image to
2289 use, this optional command may specify a path to a ramdisk
2290 (initrd) image file. If the image file will be transferred by
2291 HTCondor, then the value should just be the simple file name
2292 (no path information).
2293 xen_kernel = <included | path-to-kernel>
2296 A value of included specifies that the kernel is included in
2297 the disk file. If not one of these values, then the value is
2298 a path and file name of the kernel to be used. If a kernel
2299 file will be transferred by HTCondor, then the value should
2300 just be the simple file name (no path information).
2301 xen_kernel_params = <string>
2303 A string that is appended to the Xen kernel command line.
2304 xen_root = <string>
2307 A string that is appended to the Xen kernel command line to
2308 specify the root device. This string is required when
2309 xen_kernel gives a path to a kernel. Omission for this re‐
2310 quired case results in an error message during submission.
2311 COMMANDS FOR THE DOCKER UNIVERSE
2313 docker_image = < image-name >
2316 Defines the name of the Docker image that is the basis for
2317 the docker container.
2318 ADVANCED COMMANDS
2320 accounting_group = <accounting-group-name>
2322 Causes jobs to negotiate under the given accounting group.
2323 This value is advertised in the job ClassAd as AcctGroup. The
2324 HTCondor Administrator's manual contains more information
2325 about accounting groups.
2326 accounting_group_user = <accounting-group-user-name>
2329 Sets the user name associated with the accounting group name
2330 for resource usage accounting purposes. If not set, defaults
2331 to the value of the job ClassAd attribute Owner. This value
2332 is advertised in the job ClassAd as AcctGroupUser. If an ac‐
2333 counting group has not been set with the accounting_group
2334 command, this command is ignored.
2335 concurrency_limits = <string-list>
2338 A list of resources that this job needs. The resources are
2339 presumed to have concurrency limits placed upon them, thereby
2340 limiting the number of concurrent jobs in execution which
2341 need the named resource. Commas and space characters delimit
2342 the items in the list. Each item in the list is a string
2343 that identifies the limit, or it is a ClassAd expression that
2344 evaluates to a string, and it is evaluated in the context of
2345 machine ClassAd being considered as a match. Each item in the
2346 list also may specify a numerical value identifying the inte‐
2347 ger number of resources required for the job. The syntax
2348 follows the resource name by a colon character (:) and the
2349 numerical value. Details on concurrency limits are in the HT‐
2350 Condor Administrator's manual.
2351 concurrency_limits_expr = <ClassAd String Expression>
2354 A ClassAd expression that represents the list of resources
2355 that this job needs after evaluation. The ClassAd expression
2356 may specify machine ClassAd attributes that are evaluated
2357 against a matched machine. After evaluation, the list sets
2358 concurrency_limits.
2359 copy_to_spool = <True | False>
2362 If copy_to_spool is True, then condor_submit copies the exe‐
2363 cutable to the local spool directory before running it on a
2364 remote host. As copying can be quite time consuming and un‐
2365 necessary, the default value is False for all job universes
2366 other than the standard universe. When False, condor_submit
2367 does not copy the executable to a local spool directory. The
2368 default is True in standard universe, because resuming execu‐
2369 tion from a checkpoint can only be guaranteed to work using
2370 precisely the same executable that created the checkpoint.
2371 coresize = <size>
2373 Should the user's program abort and produce a core file,
2374 coresize specifies the maximum size in bytes of the core file
2375 which the user wishes to keep. If coresize is not specified
2376 in the command file, the system's user resource limit core‐
2377 dumpsize is used (note that coredumpsize is not an HTCondor
2378 parameter - it is an operating system parameter that can be
2379 viewed with the limit or ulimit command on Unix and in the
2380 Registry on Windows). A value of -1 results in no limits be‐
2381 ing applied to the core file size. If HTCondor is running as
2382 root, a coresize setting greater than the system coredumpsize
2383 limit will override the system setting; if HTCondor is not
2384 running as root, the system coredumpsize limit will override
2385 coresize.
2386 cron_day_of_month = <Cron-evaluated Day>
2389 The set of days of the month for which a deferral time ap‐
2390 plies. The HTCondor User's manual section on Time Scheduling
2391 for Job Execution has further details.
2392 cron_day_of_week = <Cron-evaluated Day>
2395 The set of days of the week for which a deferral time ap‐
2396 plies. The HTCondor User's manual section on Time Scheduling
2397 for Job Execution has further details.
2398 cron_hour = <Cron-evaluated Hour>
2400 The set of hours of the day for which a deferral time ap‐
2401 plies. The HTCondor User's manual section on Time Scheduling
2402 for Job Execution has further details.
2403 cron_minute = <Cron-evaluated Minute>
2405 The set of minutes within an hour for which a deferral time
2406 applies. The HTCondor User's manual section on Time Schedul‐
2407 ing for Job Execution has further details.
2408 cron_month = <Cron-evaluated Month>
2411 The set of months within a year for which a deferral time ap‐
2412 plies. The HTCondor User's manual section on Time Scheduling
2413 for Job Execution has further details.
2414 cron_prep_time = <ClassAd Integer Expression>
2417 Analogous to deferral_prep_time . The number of seconds
2418 prior to a job's deferral time that the job may be matched
2419 and sent to an execution machine.
2420 cron_window = <ClassAd Integer Expression>
2423 Analogous to the submit command deferral_window . It allows
2424 cron jobs that miss their deferral time to begin execution.
2425 The HTCondor User's manual section on Time Scheduling for Job
2427 Execution has further details.
2428 dagman_log = <pathname>
2431 DAGMan inserts this command to specify an event log that it
2432 watches to maintain the state of the DAG. If the log com‐
2433 mand is not specified in the submit file, DAGMan uses the log
2434 command to specify the event log.
2435 deferral_prep_time = <ClassAd Integer Expression>
2437 The number of seconds prior to a job's deferral time that the
2438 job may be matched and sent to an execution machine.
2439 The HTCondor User's manual section on Time Scheduling for Job
2441 Execution has further details.
2442 deferral_time = <ClassAd Integer Expression>
2445 Allows a job to specify the time at which its execution is to
2446 begin, instead of beginning execution as soon as it arrives
2447 at the execution machine. The deferral time is an expression
2448 that evaluates to a Unix Epoch timestamp (the number of sec‐
2449 onds elapsed since 00:00:00 on January 1, 1970, Coordinated
2450 Universal Time). Deferral time is evaluated with respect to
2451 the execution machine. This option delays the start of execu‐
2452 tion, but not the matching and claiming of a machine for the
2453 job. If the job is not available and ready to begin execution
2454 at the deferral time, it has missed its deferral time. A job
2455 that misses its deferral time will be put on hold in the
2456 queue.
2457 The HTCondor User's manual section on Time Scheduling for Job
2459 Execution has further details.
2460 Due to implementation details, a deferral time may not be
2462 used for scheduler universe jobs.
2463 deferral_window = <ClassAd Integer Expression>
2466 The deferral window is used in conjunction with the defer‐
2467 ral_time command to allow jobs that miss their deferral time
2468 to begin execution.
2469 The HTCondor User's manual section on Time Scheduling for Job
2471 Execution has further details.
2472 description = <string>
2475 A string that sets the value of the job ClassAd attribute
2476 JobDescription. When set, tools which display the executable
2477 such as condor_q will instead use this string.
2478 email_attributes = <list-of-job-ad-attributes>
2481 A comma-separated list of attributes from the job ClassAd.
2482 These attributes and their values will be included in the
2483 e-mail notification of job completion.
2484 image_size = <size>
2487 Advice to HTCondor specifying the maximum virtual image size
2488 to which the job will grow during its execution. HTCondor
2489 will then execute the job only on machines which have enough
2490 resources, (such as virtual memory), to support executing the
2491 job. If not specified, HTCondor will automatically make a
2492 (reasonably accurate) estimate about the job's size and ad‐
2493 just this estimate as the program runs. If specified and un‐
2494 derestimated, the job may crash due to the inability to ac‐
2495 quire more address space; for example, if malloc() fails. If
2496 the image size is overestimated, HTCondor may have difficulty
2497 finding machines which have the required resources. size is
2498 specified in KiB. For example, for an image size of 8 MiB,
2499 size should be 8000.
2500 initialdir = <directory-path>
2502 Used to give jobs a directory with respect to file input and
2503 output. Also provides a directory (on the machine from which
2504 the job is submitted) for the job event log, when a full path
2505 is not specified.
2506 For vanilla universe jobs where there is a shared file sys‐
2508 tem, it is the current working directory on the machine where
2509 the job is executed.
2510 For vanilla or grid universe jobs where file transfer mecha‐
2512 nisms are utilized (there is not a shared file system), it is
2513 the directory on the machine from which the job is submitted
2514 where the input files come from, and where the job's output
2515 files go to.
2516 For standard universe jobs, it is the directory on the ma‐
2518 chine from which the job is submitted where the condor_shadow
2519 daemon runs; the current working directory for file input and
2520 output accomplished through remote system calls.
2521 For scheduler universe jobs, it is the directory on the ma‐
2523 chine from which the job is submitted where the job runs; the
2524 current working directory for file input and output with re‐
2525 spect to relative path names.
2526 Note that the path to the executable is not relative to ini‐
2528 tialdir ; if it is a relative path, it is relative to the
2529 directory in which the condor_submit command is run.
2530 job_ad_information_attrs = <attribute-list>
2533 A comma-separated list of job ClassAd attribute names. The
2534 named attributes and their values are written to the job
2535 event log whenever any event is being written to the log.
2536 This implements the same thing as the configuration variable
2537 EVENT_LOG_INFORMATION_ATTRS (see the admin-manual/configura‐
2538 tion-macros:daemon logging configuration file entries page),
2539 but it applies to the job event log, instead of the system
2540 event log.
2541 JobBatchName = <batch_name>
2544 Set the batch name for this submit. The batch name is dis‐
2545 played by condor_q -batch. It is intended for use by users to
2546 give meaningful names to their jobs and to influence how con‐
2547 dor_q groups jobs for display. This value in a submit file
2548 can be overridden by specifying the -batch-name argument on
2549 the condor_submit command line.
2550 job_lease_duration = <number-of-seconds>
2553 For vanilla, parallel, VM, and java universe jobs only, the
2554 duration in seconds of a job lease. The default value is
2555 2,400, or forty minutes. If a job lease is not desired, the
2556 value can be explicitly set to 0 to disable the job lease se‐
2557 mantics. The value can also be a ClassAd expression that
2558 evaluates to an integer. The HTCondor User's manual section
2559 on Special Environment Considerations has further details.
2560 job_machine_attrs = <attr1, attr2, ...>
2563 A comma and/or space separated list of machine attribute
2564 names that should be recorded in the job ClassAd in addition
2565 to the ones specified by the condor_schedd daemon's system
2566 configuration variable SYSTEM_JOB_MACHINE_ATTRS
2567 . When there are multiple run attempts, history of machine
2568 attributes from previous run attempts may be kept. The number
2569 of run attempts to store may be extended beyond the sys‐
2570 tem-specified history length by using the submit file command
2571 job_machine_attrs_history_length . A machine attribute
2572 named X will be inserted into the job ClassAd as an attribute
2573 named MachineAttrX0. The previous value of this attribute
2574 will be named MachineAttrX1, the previous to that will be
2575 named MachineAttrX2, and so on, up to the specified history
2576 length. A history of length 1 means that only MachineAttrX0
2577 will be recorded. The value recorded in the job ClassAd is
2578 the evaluation of the machine attribute in the context of the
2579 job ClassAd when the condor_schedd daemon initiates the start
2580 up of the job. If the evaluation results in an Undefined or
2581 Error result, the value recorded in the job ad will be Unde‐
2582 fined or Error, respectively.
2583 want_graceful_removal = <boolean expression>
2586 If true, this job will be given a chance to shut down cleanly
2587 when removed. The job will be given as much time as the ad‐
2588 ministrator of the execute resource allows, which may be
2589 none. The default is false. For details, see the configura‐
2590 tion setting GRACEFULLY_REMOVE_JOBS.
2591 kill_sig = <signal-number>
2594 When HTCondor needs to kick a job off of a machine, it will
2595 send the job the signal specified by signal-number . sig‐
2596 nal-number needs to be an integer which represents a valid
2597 signal on the execution machine. For jobs submitted to the
2598 standard universe, the default value is the number for SIGT‐
2599 STP which tells the HTCondor libraries to initiate a check‐
2600 point of the process. For jobs submitted to other universes,
2601 the default value, when not defined, is SIGTERM, which is the
2602 standard way to terminate a program in Unix.
2603 kill_sig_timeout = <seconds>
2605 This submit command should no longer be used as of HTCondor
2606 version 7.7.3; use job_max_vacate_time instead. If
2607 job_max_vacate_time is not defined, this defines the number
2608 of seconds that HTCondor should wait following the sending of
2609 the kill signal defined by kill_sig and forcibly killing
2610 the job. The actual amount of time between sending the signal
2611 and forcibly killing the job is the smallest of this value
2612 and the configuration variable KILLING_TIMEOUT
2613 , as defined on the execute machine.
2614 load_profile = <True | False>
2617 When True, loads the account profile of the dedicated run ac‐
2618 count for Windows jobs. May not be used with run_as_owner .
2619 match_list_length = <integer value>
2622 Defaults to the value zero (0). When match_list_length is de‐
2623 fined with an integer value greater than zero (0), attributes
2624 are inserted into the job ClassAd. The maximum number of at‐
2625 tributes defined is given by the integer value. The job Clas‐
2626 sAds introduced are given as
2627 LastMatchName0 = "most-recent-Name"
2629 LastMatchName1 = "next-most-recent-Name"
2630 The value for each introduced ClassAd is given by the value
2632 of the Name attribute from the machine ClassAd of a previous
2633 execution (match). As a job is matched, the definitions for
2634 these attributes will roll, with LastMatchName1 becoming
2635 LastMatchName2, LastMatchName0 becoming LastMatchName1, and
2636 LastMatchName0 being set by the most recent value of the Name
2637 attribute.
2638 An intended use of these job attributes is in the require‐
2640 ments expression. The requirements can allow a job to prefer
2641 a match with either the same or a different resource than a
2642 previous match.
2643 job_max_vacate_time = <integer expression>
2647 An integer-valued expression (in seconds) that may be used to
2648 adjust the time given to an evicted job for gracefully shut‐
2649 ting down. If the job's setting is less than the machine's,
2650 the job's is used. If the job's setting is larger than the
2651 machine's, the result depends on whether the job has any ex‐
2652 cess retirement time. If the job has more retirement time
2653 left than the machine's max vacate time setting, then retire‐
2654 ment time will be converted into vacating time, up to the
2655 amount requested by the job.
2656 Setting this expression does not affect the job's resource
2658 requirements or preferences. For a job to only run on a ma‐
2659 chine with a minimum MachineMaxVacateTime, or to preferen‐
2660 tially run on such machines, explicitly specify this in the
2661 requirements and/or rank expressions.
2662 max_job_retirement_time = <integer expression>
2665 An integer-valued expression (in seconds) that does nothing
2666 unless the machine that runs the job has been configured to
2667 provide retirement time. Retirement time is a grace period
2668 given to a job to finish when a resource claim is about to be
2669 preempted. The default behavior in many cases is to take as
2670 much retirement time as the machine offers, so this command
2671 will rarely appear in a submit description file.
2672 When a resource claim is to be preempted, this expression in
2674 the submit file specifies the maximum run time of the job (in
2675 seconds, since the job started). This expression has no ef‐
2676 fect, if it is greater than the maximum retirement time pro‐
2677 vided by the machine policy. If the resource claim is not
2678 preempted, this expression and the machine retirement policy
2679 are irrelevant. If the resource claim is preempted the job
2680 will be allowed to run until the retirement time expires, at
2681 which point it is hard-killed. The job will be soft-killed
2682 when it is getting close to the end of retirement in order to
2683 give it time to gracefully shut down. The amount of lead-time
2684 for soft-killing is determined by the maximum vacating time
2685 granted to the job.
2686 Standard universe jobs and any jobs running with nice_user
2688 priority have a default max_job_retirement_time of 0, so no
2689 retirement time is utilized by default. In all other cases,
2690 no default value is provided, so the maximum amount of re‐
2691 tirement time is utilized by default.
2692 Setting this expression does not affect the job's resource
2694 requirements or preferences. For a job to only run on a ma‐
2695 chine with a minimum MaxJobRetirementTime, or to preferen‐
2696 tially run on such machines, explicitly specify this in the
2697 requirements and/or rank expressions.
2698 nice_user = <True | False>
2700 Normally, when a machine becomes available to HTCondor, HT‐
2701 Condor decides which job to run based upon user and job pri‐
2702 orities. Setting nice_user equal to True tells HTCondor not
2703 to use your regular user priority, but that this job should
2704 have last priority among all users and all jobs. So jobs sub‐
2705 mitted in this fashion run only on machines which no other
2706 non-nice_user job wants - a true bottom-feeder job! This is
2707 very handy if a user has some jobs they wish to run, but do
2708 not wish to use resources that could instead be used to run
2709 other people's HTCondor jobs. Jobs submitted in this fashion
2710 have "nice-user." prepended to the owner name when viewed
2711 from condor_q or condor_userprio. The default value is False.
2712 noop_job = <ClassAd Boolean Expression>
2714 When this boolean expression is True, the job is immediately
2715 removed from the queue, and HTCondor makes no attempt at run‐
2716 ning the job. The log file for the job will show a job sub‐
2717 mitted event and a job terminated event, along with an exit
2718 code of 0, unless the user specifies a different signal or
2719 exit code.
2720 noop_job_exit_code = <return value>
2723 When noop_job is in the submit description file and evalu‐
2724 ates to True, this command allows the job to specify the re‐
2725 turn value as shown in the job's log file job terminated
2726 event. If not specified, the job will show as having termi‐
2727 nated with status 0. This overrides any value specified with
2728 noop_job_exit_signal .
2729 noop_job_exit_signal = <signal number>
2732 When noop_job is in the submit description file and evalu‐
2733 ates to True, this command allows the job to specify the sig‐
2734 nal number that the job's log event will show the job having
2735 terminated with.
2736 remote_initialdir = <directory-path>
2739 The path specifies the directory in which the job is to be
2740 executed on the remote machine. This is currently supported
2741 in all universes except for the standard universe.
2742 rendezvousdir = <directory-path>
2745 Used to specify the shared file system directory to be used
2746 for file system authentication when submitting to a remote
2747 scheduler. Should be a path to a preexisting directory.
2748 run_as_owner = <True | False>
2751 A boolean value that causes the job to be run under the login
2752 of the submitter, if supported by the joint configuration of
2753 the submit and execute machines. On Unix platforms, this de‐
2754 faults to True, and on Windows platforms, it defaults to
2755 False. May not be used with load_profile . See the HTCondor
2756 manual Platform-Specific Information chapter for administra‐
2757 tive details on configuring Windows to support this option.
2758 stack_size = <size in bytes>
2760 This command applies only to Linux platform jobs that are not
2761 standard universe jobs. An integer number of bytes, repre‐
2762 senting the amount of stack space to be allocated for the
2763 job. This value replaces the default allocation of stack
2764 space, which is unlimited in size.
2765 submit_event_notes = <note>
2767 A string that is appended to the submit event in the job's
2768 log file. For DAGMan jobs, the string DAG Node: and the
2769 node's name is automatically defined for submit_event_notes,
2770 causing the logged submit event to identify the DAG node job
2771 submitted.
2772 +<attribute> = <value>
2774 A line that begins with a '+' (plus) character instructs con‐
2775 dor_submit to insert the given attribute into the job ClassAd
2776 with the given value. Note that setting an attribute should
2777 not be used in place of one of the specific commands listed
2778 above. Often, the command name does not directly correspond
2779 to an attribute name; furthermore, many submit commands re‐
2780 sult in actions more complex than simply setting an attribute
2781 or attributes. See /classad-attributes/job-classad-attributes
2782 for a list of HTCondor job attributes.
2783 MACROS AND COMMENTS
2785 In addition to commands, the submit description file can contain macros
2788 and comments.
2789 Macros Parameterless macros in the form of $(macro_name:default ini‐
2791 tial value) may be used anywhere in HTCondor submit descrip‐
2792 tion files to provide textual substitution at submit time.
2793 Macros can be defined by lines in the form of
2794 <macro_name> = <string>
2796 Two pre-defined macros are supplied by the submit description
2798 file parser. The $(Cluster) or $(ClusterId) macro supplies
2799 the value of the
2800 ClusterId job ClassAd attribute, and the $(Process) or
2802 $(ProcId) macro supplies the value of the ProcId job ClassAd
2803 attribute. These macros are intended to aid in the specifica‐
2804 tion of input/output files, arguments, etc., for clusters
2805 with lots of jobs, and/or could be used to supply an HTCondor
2806 process with its own cluster and process numbers on the com‐
2807 mand line.
2808 The $(Node) macro is defined for parallel universe jobs, and
2810 is especially relevant for MPI applications. It is a unique
2811 value assigned for the duration of the job that essentially
2812 identifies the machine (slot) on which a program is execut‐
2813 ing. Values assigned start at 0 and increase monotonically.
2814 The values are assigned as the parallel job is about to
2815 start.
2816 Recursive definition of macros is permitted. An example of a
2818 construction that works is the following:
2819 foo = bar
2821 foo = snap $(foo)
2822 As a result, foo = snap bar.
2824 Note that both left- and right- recursion works, so
2826 foo = bar
2828 foo = $(foo) snap
2829 has as its result foo = bar snap.
2831 The construction
2833 foo = $(foo) bar
2835 by itself will not work, as it does not have an initial base
2837 case. Mutually recursive constructions such as:
2838 B = bar
2840 C = $(B)
2841 B = $(C) boo
2842 will not work, and will fill memory with expansions.
2844 A default value may be specified, for use if the macro has no
2846 definition. Consider the example
2847 D = $(E:24)
2849 Where E is not defined within the submit description file,
2851 the default value 24 is used, resulting in
2852 D = 24
2854 This is of limited value, as the scope of macro substitution
2856 is the submit description file. Thus, either the macro is or
2857 is not defined within the submit description file. If the
2858 macro is defined, then the default value is useless. If the
2859 macro is not defined, then there is no point in using it in a
2860 submit command.
2861 To use the dollar sign character ($) as a literal, without
2864 macro expansion, use
2865 $(DOLLAR)
2867 In addition to the normal macro, there is also a special kind
2869 of macro called a substitution macro
2870 that allows the substitution of a machine ClassAd attribute
2871 value defined on the resource machine itself (gotten after a
2872 match to the machine has been made) into specific commands
2873 within the submit description file. The substitution macro is
2874 of the form:
2875 $$(attribute)
2877 As this form of the substitution macro is only evaluated
2879 within the context of the machine ClassAd, use of a scope
2880 resolution prefix TARGET. or MY. is not allowed.
2881 A common use of this form of the substitution macro is for
2883 the heterogeneous submission of an executable:
2884 executable = povray.$$(OpSys).$$(Arch)
2886 Values for the OpSys and Arch attributes are substituted at
2888 match time for any given resource. This example allows HTCon‐
2889 dor to automatically choose the correct executable for the
2890 matched machine.
2891 An extension to the syntax of the substitution macro provides
2893 an alternative string to use if the machine attribute within
2894 the substitution macro is undefined. The syntax appears as:
2895 $$(attribute:string_if_attribute_undefined)
2897 An example using this extended syntax provides a path name to
2899 a required input file. Since the file can be placed in dif‐
2900 ferent locations on different machines, the file's path name
2901 is given as an argument to the program.
2902 arguments = $$(input_file_path:/usr/foo)
2904 On the machine, if the attribute input_file_path is not de‐
2906 fined, then the path /usr/foo is used instead.
2907 A further extension to the syntax of the substitution macro
2909 allows the evaluation of a ClassAd expression to define the
2910 value. In this form, the expression may refer to machine at‐
2911 tributes by prefacing them with the TARGET. scope resolution
2912 prefix. To place a ClassAd expression into the substitution
2913 macro, square brackets are added to delimit the expression.
2914 The syntax appears as:
2915 $$([ClassAd expression])
2917 An example of a job that uses this syntax may be one that
2919 wants to know how much memory it can use. The application
2920 cannot detect this itself, as it would potentially use all of
2921 the memory on a multi-slot machine. So the job determines the
2922 memory per slot, reducing it by 10% to account for miscella‐
2923 neous overhead, and passes this as a command line argument to
2924 the application. In the submit description file will be
2925 arguments = --memory $$([TARGET.Memory * 0.9])
2927 To insert two dollar sign characters ($$) as literals into a
2931 ClassAd string, use
2932 $$(DOLLARDOLLAR)
2934 The environment macro, $ENV, allows the evaluation of an en‐
2940 vironment variable to be used in setting a submit description
2941 file command. The syntax used is
2942 $ENV(variable)
2944 An example submit description file command that uses this
2946 functionality evaluates the submitter's home directory in or‐
2947 der to set the path and file name of a log file:
2948 log = $ENV(HOME)/jobs/logfile
2950 The environment variable is evaluated when the submit de‐
2952 scription file is processed.
2953 The $RANDOM_CHOICE macro allows a random choice to be made
2958 from a given list of parameters at submission time. For an
2959 expression, if some randomness needs to be generated, the
2960 macro may appear as
2961 $RANDOM_CHOICE(0,1,2,3,4,5,6)
2963 When evaluated, one of the parameters values will be chosen.
2965 Comments
2967 Blank lines and lines beginning with a pound sign ('#') char‐
2968 acter are ignored by the submit description file parser.
2969
2971 While processing the queue command in a submit file or from the command
2972 line, condor_submit will set the values of several automatic submit
2973 variables so that they can be referred to by statements in the submit
2974 file. With the exception of Cluster and Process, if these variables are
2975 set by the submit file, they will not be modified during queue pro‐
2976 cessing.
2977 ClusterId
2979 Set to the integer value that the ClusterId attribute that
2980 the job ClassAd will have when the job is submitted. All jobs
2981 in a single submit will normally have the same value for the
2982 ClusterId. If the -dry-run argument is specified, The value
2983 will be 1.
2984 Cluster
2986 Alternate name for the ClusterId submit variable. Before HT‐
2987 Condor version 8.4 this was the only name.
2988 ProcId Set to the integer value that the ProcId attribute of the job
2990 ClassAd will have when the job is submitted. The value will
2991 start at 0 and increment by 1 for each job submitted.
2992 Process
2994 Alternate name for the ProcId submit variable. Before HTCon‐
2995 dor version 8.4 this was the only name.
2996 Node For parallel universes, set to the value #pArAlLeLnOdE# or
2998 #MpInOdE# depending on the parallel universe type For other
2999 universes it is set to nothing.
3000 Step Set to the step value as it varies from 0 to N-1 where N is
3002 the number provided on the queue argument. This variable
3003 changes at the same rate as ProcId when it changes at all.
3004 For submit files that don't make use of the queue number op‐
3005 tion, Step will always be 0. For submit files that don't make
3006 use of any of the foreach options, Step and ProcId will al‐
3007 ways be the same.
3008 ItemIndex
3010 Set to the index within the item list being processed by the
3011 various queue foreach options. For submit files that don't
3012 make use of any queue foreach list, ItemIndex will always be
3013 0 For submit files that make use of a slice to select only
3014 some items in a foreach list, ItemIndex will only be set to
3015 selected values.
3016 Row Alternate name for ItemIndex.
3018 Item when a queue foreach option is used and no variable list is
3020 supplied, this variable will be set to the value of the cur‐
3021 rent item.
3022 The automatic variables below are set before parsing the submit file,
3024 and will not vary during processing unless the submit file itself sets
3025 them.
3026 ARCH Set to the CPU architecture of the machine running con‐
3028 dor_submit. The value will be the same as the automatic con‐
3029 figuration variable of the same name.
3030 OPSYS Set to the name of the operating system on the machine run‐
3032 ning condor_submit. The value will be the same as the auto‐
3033 matic configuration variable of the same name.
3034 OPSYSANDVER
3036 Set to the name and major version of the operating system on
3037 the machine running condor_submit. The value will be the same
3038 as the automatic configuration variable of the same name.
3039 OPSYSMAJORVER
3041 Set to the major version of the operating system on the ma‐
3042 chine running condor_submit. The value will be the same as
3043 the automatic configuration variable of the same name.
3044 OPSYSVER
3046 Set to the version of the operating system on the machine
3047 running condor_submit. The value will be the same as the au‐
3048 tomatic configuration variable of the same name.
3049 SPOOL Set to the full path of the HTCondor spool directory. The
3051 value will be the same as the automatic configuration vari‐
3052 able of the same name.
3053 IsLinux
3055 Set to true if the operating system of the machine running
3056 condor_submit is a Linux variant. Set to false otherwise.
3057 IsWindows
3059 Set to true if the operating system of the machine running
3060 condor_submit is a Microsoft Windows variant. Set to false
3061 otherwise.
3062 SUBMIT_FILE
3064 Set to the full pathname of the submit file being processed
3065 by condor_submit. If submit statements are read from standard
3066 input, it is set to nothing.
3067
3069 condor_submit will exit with a status value of 0 (zero) upon success,
3070 and a non-zero value upon failure.
3071
3073 • Submit Description File Example 1: This example queues three jobs for
3074 execution by HTCondor. The first will be given command line arguments
3075 of 15 and 2000, and it will write its standard output to foo.out1.
3076 The second will be given command line arguments of 30 and 2000, and
3077 it will write its standard output to foo.out2. Similarly the third
3078 will have arguments of 45 and 6000, and it will use foo.out3 for its
3079 standard output. Standard error output (if any) from all three pro‐
3080 grams will appear in foo.error.
3081 ####################
3083 #
3084 # submit description file
3085 # Example 1: queuing multiple jobs with differing
3086 # command line arguments and output files.
3087 #
3088 ####################
3089 Executable = foo
3091 Universe = vanilla
3092 Arguments = 15 2000
3094 Output = foo.out0
3095 Error = foo.err0
3096 Queue
3097 Arguments = 30 2000
3099 Output = foo.out1
3100 Error = foo.err1
3101 Queue
3102 Arguments = 45 6000
3104 Output = foo.out2
3105 Error = foo.err2
3106 Queue
3107 Or you can get the same results as the above submit file by using a
3109 list of arguments with the Queue statement
3110 ####################
3112 #
3113 # submit description file
3114 # Example 1b: queuing multiple jobs with differing
3115 # command line arguments and output files, alternate syntax
3116 #
3117 ####################
3118 Executable = foo
3120 Universe = vanilla
3121 # generate different output and error filenames for each process
3123 Output = foo.out$(Process)
3124 Error = foo.err$(Process)
3125 Queue Arguments From (
3127 15 2000
3128 30 2000
3129 45 6000
3130 )
3131 • Submit Description File Example 2: This submit description file exam‐
3133 ple queues 150 runs of program foo which must have been compiled and
3134 linked for an Intel x86 processor running RHEL 3. HTCondor will not
3135 attempt to run the processes on machines which have less than 32
3136 Megabytes of physical memory, and it will run them on machines which
3137 have at least 64 Megabytes, if such machines are available. Stdin,
3138 stdout, and stderr will refer to in.0, out.0, and err.0 for the first
3139 run of this program (process 0). Stdin, stdout, and stderr will refer
3140 to in.1, out.1, and err.1 for process 1, and so forth. A log file
3141 containing entries about where and when HTCondor runs, takes check‐
3142 points, and migrates processes in this cluster will be written into
3143 file foo.log.
3144 ####################
3146 #
3147 # Example 2: Show off some fancy features including
3148 # use of pre-defined macros and logging.
3149 #
3150 ####################
3151 Executable = foo
3153 Universe = standard
3154 Requirements = OpSys == "LINUX" && Arch =="INTEL"
3155 Rank = Memory >= 64
3156 Request_Memory = 32 Mb
3157 Image_Size = 28 Mb
3158 Error = err.$(Process)
3160 Input = in.$(Process)
3161 Output = out.$(Process)
3162 Log = foo.log
3163 Queue 150
3164 • Submit Description File Example 3: This example targets the
3166 /bin/sleep program to run only on a platform running a RHEL 6 operat‐
3167 ing system. The example presumes that the pool contains machines run‐
3168 ning more than one version of Linux, and this job needs the particu‐
3169 lar operating system to run correctly.
3170 ####################
3172 #
3173 # Example 3: Run on a RedHat 6 machine
3174 #
3175 ####################
3176 Universe = vanilla
3177 Executable = /bin/sleep
3178 Arguments = 30
3179 Requirements = (OpSysAndVer == "RedHat6")
3180 Error = err.$(Process)
3182 Input = in.$(Process)
3183 Output = out.$(Process)
3184 Log = sleep.log
3185 Queue
3186 • Command Line example: The following command uses the -append option
3188 to add two commands before the job(s) is queued. A log file and an
3189 error log file are specified. The submit description file is un‐
3190 changed.
3191 condor_submit -a "log = out.log" -a "error = error.log" mysubmitfile
3193 Note that each of the added commands is contained within quote marks
3195 because there are space characters within the command.
3196 • periodic_remove example: A job should be removed from the queue, if
3198 the total suspension time of the job is more than half of the run
3199 time of the job.
3200 Including the command
3202 periodic_remove = CumulativeSuspensionTime >
3204 ((RemoteWallClockTime - CumulativeSuspensionTime) / 2.0)
3205 in the submit description file causes this to happen.
3207
3209 • For security reasons, HTCondor will refuse to run any jobs submitted
3210 by user root (UID = 0) or by a user whose default group is group
3211 wheel (GID = 0). Jobs submitted by user root or a user with a default
3212 group of wheel will appear to sit forever in the queue in an idle
3213 state.
3214 • All path names specified in the submit description file must be less
3216 than 256 characters in length, and command line arguments must be
3217 less than 4096 characters in length; otherwise, condor_submit gives a
3218 warning message but the jobs will not execute properly.
3219 • Somewhat understandably, behavior gets bizarre if the user makes the
3221 mistake of requesting multiple HTCondor jobs to write to the same
3222 file, and/or if the user alters any files that need to be accessed by
3223 an HTCondor job which is still in the queue. For example, the com‐
3224 pressing of data or output files before an HTCondor job has completed
3225 is a common mistake.
3226 • To disable checkpointing for Standard Universe jobs, include the
3228 line:
3229 +WantCheckpoint = False
3231 in the submit description file before the queue command(s).
3233
3235 HTCondor User Manual
3236
3238 HTCondor Team
3239
3241 1990-2021, Center for High Throughput Computing, Computer Sciences De‐
3242 partment, University of Wisconsin-Madison, Madison, WI, US. Licensed
3243 under the Apache License, Version 2.0.
32448.8 Jan 26, 2021 CONDOR_SUBMIT(1)