1condor_annex(1) General Commands Manual condor_annex(1)
2
6 condor_annexAdd cloud resources to the pool.
7
9 condor_annex-help
10 condor_annex[-aws-region <region>] -setup
12 [/full/path/to/access/key/file [/full/path/to/secret/key/file]]
13 condor_annex[-aws-on-demand] -annex-name <name of the annex>-count
15 <integer number of instances>[-aws-on-demand-*] [common options]
16 condor_annex[-aws-spot-fleet] -annex-name <name of the annex>-slots
18 <integer weight>[-aws-spot-fleet-*] [common options]
19 condor_annex-annex-name <name of the annex>-duration hours
21 condor_annex[-annex-name <name of the annex>] -status[-classad]
23 condor_annex-check-setup
25 condor_annex<condor_annex options>status<condor_status options>
27
29 condor_annexadds clouds resources to the pool. (“The pool”
30 is determined in the usual manner for HTCondor daemons and tools.) This
31 version supports only Amazon Web Services (`AWS'). To add “on-
32 demand” instances, use the third form listed above; to add
33 “spot” instances, use the fourth. For an explanation of
34 terms, consult either the HTCondor manual (chapter ) or the AWS docu‐
35 mentation.
36 Using condor_annexwith AWS requires a one-time setup procedure per‐
38 formed by invoking condor_annexwith the -setupflag (the second form
39 listed above). You may check if this procedure has been performed with
40 the -check-setupflag (the seventh form listed above).
41 To reset the lease on an existing annex, invoke condor_annexwith only
43 the -annex-nameoption and -durationflag (the fifth form listed above).
44 To determine which of the instances previously requested for a particu‐
46 lar annex are not currently in the pool, invoke condor_annexwith the
47 -statusflag and the -annex-nameoption (the sixth form listed above).
48 The output of this command is intended to be human-readable; specifying
49 the -classadflag will produce the same information in ClassAd format.
50 If you omit -annex-name, information for all annexes will be returned.
51 Starting in 8.7.3, you may instead invoke condor_annexwith statusas a
53 command argument (the eighth form listed above). This will cause con‐
54 dor_annexto use condor_statusto present annex instance data. Arguments
55 and options on the command line after statuswill be passed unmodified
56 to condor_status, but not all arguments and options will behave as
57 expected. (See below.) condor_annexwill construct an ad for each annex
58 instance and pass that information to condor_status; condor_statuswill
59 (unless you specify otherwise using its command line) query the collec‐
60 tor for more information about the instances. Information from the col‐
61 lector will be presented as usual; instances which did not have ads in
62 the collector will be presented last, in their own table. These
63 instances can not be presented in the usual way because the annex
64 instance ads generated by condor_annexdo not (and can not) have the
65 same information in them as ads generated by a condor_startdrunning in
66 the instance. See the condor_statusdocumentation (section ) for details
67 about the “merge” mode of condor_statusused by this command
68 argument. Note that both condor_annexand condor_statushave -annex-
69 nameoptions; if you're interested in a particular annex, put this flag
70 on the command line beforethe statuscommand argument to avoid confusing
71 results.
72 Common options are listed first, followed by options specific to AWS,
74 followed by options specific to AWS' on-demand instances, followed by
75 options specific to AWS' spot instances, followed by options intended
76 for use by experts.
77
79 -help
80 Print a usage reminder.
82 -setup [/full/path/to/access/key/file /full/path/to/secret/key/file]
88 Do the first-time setup.
90 -duration hours
96 Set the maximum lease duration in decimal hours. After this amount
98 of time, all instances will terminated, regardless of their idle‐
99 ness. Defaults to 50 minutes.
100 -idle hours
104 Set the maximum idle duration in decimal hours. An instance idle for
106 longer than this duration will terminate itself. Defaults to 15 min‐
107 utes.
108 -config-dir /full/path/to/directory
112 Copy the contents of /full/path/to/directoryto each instance's con‐
114 figuration directory.
115 -owner owner[, owner]*
119 Configure the annex so that only ownermay start jobs there. By
121 default, configure the annex so that only the user running con‐
122 dor_annexmay start jobs there.
123 -no-owner
127 Configure the annex so that anyone in the pool may use the annex.
129 -aws-region region
135 Specify the region in which to create the annex.
137 -aws-user-data user-data
141 Set the instance user data to user-data.
143 -aws-user-data-file /full/path/to/file
147 Set the instance user data to the contents of the file
149 /full/path/to/file.
150 -aws-default-user-data user-data
154 Set the instance user data to user-data, if it's not already set.
156 Only applies to spot fleet requests.
157 -aws-default-user-data-file /full/path/to/file
161 Set the instance user data to the contents of the file
163 /full/path/to/file, if it's not already set. Only applies to spot
164 fleet requests.
165 -aws-on-demand-instance-type instance-type
171 This annex will requests instances of type instance-type. The
173 default for v8.7.1 is `m4.large'.
174 -aws-on-demand-ami-id ami-id
178 This annex will start instances of the AMI ami-id. The default for
180 v8.7.1 is `ami-35b13223', a GPU-compatible Amazon Linux image with
181 HTCondor pre-installed.
182 -aws-on-demand-security-group-ids group-id[,group-id]
186 This annex will start instances with the listed security group IDs.
188 The default is the security group created by -setup.
189 -aws-on-demand-key-name key-name
193 This annex will start instances with the key pair named key-name.
195 The default is the key pair created by -setup.
196 -aws-spot-fleet-config-file /full/path/to/file
202 Use the JSON blob in /full/path/to/filefor the spot fleet request.
204 -aws-access-key-file /full/path/to/access-key-file
210 Experts only.
212 -aws-secret-key-file /full/path/to/secret-key-file
216 Experts only.
218 -aws-ec2-url https://ec2.<region>.amazonaws.com
222 Experts only.
224 -aws-events-url https://events.<region>.amazonaws.com
228 Experts only.
230 -aws-lambda-url https://lambda.<region>.amazonaws.com
234 Experts only.
236 -aws-s3-url https://s3.<region>.amazonaws.com
240 Experts only.
242 -aws-spot-fleet-lease-function-arn sfr-lease-function-arn
246 Developers only.
248 -aws-on-demand-lease-function-arn odi-lease-function-arn
252 Developers only.
254 -aws-on-demand-instance-profile-arn instance-profile-arn
258 Developers only.
260
264 Currently, only AWS is supported. The AMI configured by setup runs
265 HTCondor v8.6.10 on Amazon Linux 2016.09, and the default instance type
266 is “m4.large”. The default AMI has the appropriate drivers
267 for AWS' GPU instance types.
268
270 To start an on-demand annex named `MyFirstAnnex' with one core, using
271 the default AMI and instance type, run
272 condor_annex -count 1 -annex-name MyFirstAnnex
274 You will be asked to confirm that the defaults are what you want.
276 As of 2017-04-17, the following example will cost a minimum of $90.
278 To start an on-demand annex with 100 GPUs that job owners `big' and
280 `little' may use (be sure to include yourself!), run
281 condor_annex -count 100 -annex-name MySecondAnnex \
283 -aws-on-demand-instance-type p2.xlarge -owner "big, little"
284
286 condor_annexwill exit with a status value of 0 (zero) on success.
287
289 Center for High Throughput Computing, University of Wiscon‐
290 sin–Madison
291
293 Copyright © 1990-2019 Center for High Throughput Computing, Computer
294 Sciences Department, University of Wisconsin-Madison, Madison, WI. All
295 Rights Reserved. Licensed under the Apache License, Version 2.0.
296 date condor_annex(1)