1JobManager(3pm) User Contributed Perl Documentation JobManager(3pm)
2
6 Globus::GRAM::JobManager - Base class for all Job Manager scripts
7
9 $manager = new Globus::GRAM::JobManager($job_description);
10 $manager->log("Starting new operation");
12 $manager->nfssync($fileobj,$createflag);
13 $manager->respond($hashref);
14 $hashref = $manager->submit();
15 $hashref = $manager->poll();
16 $hashref = $manager->cancel();
17 $hashref = $manager->signal();
18 $hashref = $manager->make_scratchdir();
19 $hashref = $manager->remove_scratchdir();
20 $hashref = $manager->rewrite_urls();
21 $hashref = $manager->stage_in();
22 $hashref = $manager->stage_out();
23 $hashref = $manager->cache_cleanup();
24 $hashref = $manager->remote_io_file_create();
25 $hashref = $manager->proxy_relocate();
26 $hashref = $manager->proxy_update();
27 $scalar = $manager->pipe_out_cmd(@arglist);
28 ($stderr, $rc) = $manager->pipe_err_cmd(@arglist);
29 $status = $manager->fork_and_exec_cmd(@arglist);
30 $manager->append_path($hash, $variable, $path);
31 $scalar = $manager->setup_softenv();
32
34 The Globus::GRAM::JobManager module implements the base behavior for a
35 Job Manager script interface. Scheduler-specific job manager scripts
36 must inherit from this module in order to be used by the job manager.
37 Methods
39 $manager = Globus::GRAM::JobManager->new($JobDescription)
40 Each Globus::GRAM::JobManager object is created by calling the
41 constructor with a single argument, a Globus::GRAM::JobDescription
42 object containing the information about the job request which the
43 script will be modifying. Modules which subclass
44 Globus::GRAM::JobManager MUST call the super-class's constructor,
45 as in this code fragment:
46 my $proto = shift;
48 my $class = ref($proto) || $proto;
49 my $self = $class->SUPER::new(@_);
50 bless $self, $class;
52 $manager->log($string)
54 Log a message to the job manager log file. The message is preceded
55 by a timestamp.
56 $manager->nfssync($object,$create)
58 Send an NFS update by touching the file (or directory) in question.
59 If the $create is true, a file will be created. If it is false, the
60 $object will not be created.
61 $manager->respond($message)
63 Send a response to the job manager program. The response may either
64 be a hash reference consisting of a hash of (variable, value)
65 pairs, which will be returned to the job manager, or an already
66 formatted string. This only needs to be directly called by a job
67 manager implementation when the script wants to send a partial
68 response while processing one of the scheduler interface methods
69 (for example, to indicate that a file has been staged).
70 The valid keys for a response are defined in the RESPONSES section.
72 $manager->submit()
74 Submit a job request to the scheduler. The default implementation
75 returns with the Globus::GRAM::Error::UNIMPLEMENTED error.
76 Scheduler specific subclasses should reimplement this method to
77 submit the job to the scheduler.
78 A scheduler which implements this method should return a hash
80 reference containing a scheduler-specific job identifier as the
81 value of the hash's JOB_ID key, and optionally, the a GRAM job
82 state as the value of the hash's JOB_STATE key if the job
83 submission was successful; otherwise a Globus::GRAM::Error value
84 should be returned. The job state values are defined in the
85 Globus::GRAM::JobState module. The job parameters (as found in the
86 job rsl) are defined in Globus::GRAM::Jobdescription object in
87 $self->{JobDescription}.
88 For example:
90 return {JOB_STATE => Globus::GRAM::JobState::PENDING,
92 JOB_ID => $job_id};
93 $manager->poll()
95 Poll a job's status. The default implementation returns with the
96 Globus::GRAM::Error::UNIMPLEMENTED error. Scheduler specific
97 subclasses should reimplement this method to poll the scheduler.
98 A scheduler which implements this method should return a hash
100 reference containing the JOB_STATE value. The job's ID can be
101 accessed by calling the $self->{JobDescription}->jobid() method.
102 $manager->cancel()
104 Cancel a job. The default implementation returns with the
105 Globus::GRAM::Error::UNIMPLEMENTED error. Scheduler specific
106 subclasses should reimplement this method to remove the job from
107 the scheduler.
108 A scheduler which implements this method should return a hash
110 reference containing the JOB_STATE value. The job's ID can be
111 accessed by calling the $self->{JobDescription}->jobid() method.
112 $manager->signal()
114 Signal a job. The default implementation returns with the
115 Globus::GRAM::Error::UNIMPLEMENTED error. Scheduler specific
116 subclasses should reimplement this method to remove the job from
117 the scheduler. The JobManager module can determine the job's ID,
118 the signal number, and the (optional) signal arguments from the Job
119 Description by calling it's job_id(), signal(), and and
120 signal_arg() methods, respectively.
121 Depending on the signal, it may be appropriate for the JobManager
123 object to return a hash reference containing a JOB_STATE update.
124 $manager->make_scratchdir()
126 Create a scratch directory for a job. The scratch directory
127 location is based on the JobDescription's scratch_dir_base() and
128 scratch_dir() methods.
129 If the scratch_dir() value is a relative path, then a directory
131 will be created as a subdirectory of
132 scratch_dir_base()/scratch_dir(), otherwise, it will be created as
133 a subdirectory of scratch_dir(). This method will return a hash
134 reference containing mapping SCRATCH_DIR to the absolute path of
135 newly created scratch directory if successful.
136 $manager->make_scratchdir()
138 Delete a job's scratch directory. All files and subdirectories of
139 the JobDescription's scratch_directory() will be deleted.
140 $manager->make_scratchdir()
142 Delete some job-related files. All files listed in the
143 JobDescription's file_cleanup() array will be deleted.
144 $manager->rewrite_urls()
146 Looks up URLs listed in the JobDescription's stdin() and
147 executable(), and replaces them with paths to locally cached
148 copies.
149 $manager->stage_in()
151 Stage input files need for the job from remote storage. The files
152 to be staged are defined by the array of [URL, path] pairs in the
153 job description's file_stage_in() and file_stage_in_shared()
154 methods. The Globus::GRAM::JobManager module provides an
155 implementation of this functionality using the globus-url-copy and
156 globus-gass-cache programs. Files which are staged in are not
157 automatically removed when the job terminates.
158 This function returns intermediate responses using the
160 Globus::GRAM::JobManager::response() method to let the job manager
161 know when each individual file has been staged.
162 $manager->stage_out()
164 Stage output files generated by this job to remote storage. The
165 files to be staged are defined by the array of [URL, destination]
166 pairs in the job description's file_stage_out() method. The
167 Globus::GRAM::JobManager module provides an implementation of this
168 functionality using the globus-url-copy program. Files which are
169 staged out are not removed by this method.
170 $manager->cache_cleanup()
172 Clean up cache references in the GASS which match this job's cache
173 tag .
174 $manager->remote_io_file_create()
176 Create the remote I/O file in the job dir which will contain the
177 remote_io_url RSL attribute's value.
178 $manager->proxy_relocate()
180 Relocate the delegated proxy for job execution. Job Managers need
181 to override the default if they intend to relocate the proxy into
182 some common file system other than the cache. The job manager
183 program does not depend on the new location of the proxy. Job
184 Manager modules must not remove the default proxy.
185 $hashref = $manager->proxy_update();
187 $manager->append_path($ref, $var, $path)
188 Append $path to the value of $ref->{$var}, dealing with the case
189 where $ref->{$var} is not yet defined.
190 $manager->pipe_out_cmd(@arg)
192 Create a new process to run the first argument application with the
193 remaining arguments (which may be empty). No shell metacharacter
194 will be evaluated, avoiding a shell invocation. Stderr is
195 redirected to /dev/null and stdout is being captured by the parent
196 process, which is also the result returned. In list mode, all
197 lines are returned, in scalar mode, only the first line is being
198 returned. The line termination character is already cut off. Use
199 this function as more efficient backticks, if you do not need shell
200 metacharacter evaluation.
201 Caution: This function deviates in two manners from regular
203 backticks. Firstly, it chomps the line terminator from the output.
204 Secondly, it returns only the first line in scalar context instead
205 of a multiline concatinated string. As with regular backticks, the
206 result may be undefined in scalar context, if no result exists.
207 A child error code with an exit code of 127 indicates that the
209 application could not be run. The scalar result returned by this
210 function is usually undef'ed in this case.
211 ($stder, $rc) = $manager->pipe_err_cmd(@arg)
213 Create a new process to run the first argument application with the
214 remaining arguments (which may be empty). No shell metacharacter
215 will be evaluated, avoiding a shell invocation.
216 This method returns a list of two items, the standard error of the
218 program, and the exit code of the program. If the error code is
219 127, then the application could not be run. Standard output is
220 discarded.
221 $manager->fork_and_exec_cmd(@arg)
223 Fork off a child to run the first argument in the list. Remaining
224 arguments will be passed, but shell interpolation is avoided.
225 Signals SIGINT and SIGQUIT are ignored in the child process. Stdout
226 is appended to /dev/null, and stderr is dup2 from stdout. The
227 parent waits for the child to finish, and returns the value for the
228 CHILD_ERROR variable as result. Use this function as more efficient
229 system() call, if you can do not need shell metacharacter
230 evaluation.
231 Note that the inability to execute the program will result in a
233 status code of 127.
234 $manager->job_dir()
236 Return the temporary directory to store job-related files, which
237 have no need for file caching.
238 $manager->setup_softenv()
240 Either add a line to the specified command script file handle to
241 load the user's default SoftEnv configuration, or create a custom
242 SoftEnv script and add commands to the specified command script
243 file handle to load it.
244
246 When returning from a job interface method, or when sending an
247 intermediate response via the response() method, the following hash
248 keys are valid:
249 · JOB_STATE
251 An integer job state value. These are enumerated in the
253 Globus::GRAM::JobState module.
254 · ERROR
256 An integer error code. These are enumerated in the
258 Globus::GRAM::Error module.
259 · JOB_ID
261 A string containing a job identifier, which can be used to poll,
263 cancel, or signal a job in progress. This response should only be
264 returned by the submit method.
265 · SCRATCH_DIR
267 A string containing the path to a newly-created scratch directory.
269 This response should only be returned by the make_scratchdir
270 method.
271 · STAGED_IN
273 A string containing the (URL, path) pair for a file which has now
275 been staged in. This response should only be returned by the
276 stage_in method.
277 · STAGED_IN_SHARED
279 A string containing the (URL, path) pair for a file which has now
281 been staged in and symlinked from the cache. This response should
282 only be returned by the stage_in_shared method.
283 · STAGED_OUT
285 A string containing the (path, URL) pair for a file which has now
287 been staged out by the script. This response should only be
288 returned by the stage_out method.
289perl v5.16.3 2020-12-12 JobManager(3pm)