1CONDOR_VACATE_JOB(1) HTCondor Manual CONDOR_VACATE_JOB(1)
2
6 condor_vacate_job - HTCondor Manual
7 vacate jobs in the HTCondor queue from the hosts where they are running
9
12 condor_vacate_job [-help | -version ]
13 condor_vacate_job [ -pool centralmanagerhostname[:portnumber] | -name
15 scheddname ] | [-addr "<a.b.c.d:port>"] [-fast ] cluster... | clus‐
16 ter.process... | user... | -constraint expression ...
17 condor_vacate_job [ -pool centralmanagerhostname[:portnumber] | -name
19 scheddname ] | [-addr "<a.b.c.d:port>"] [-fast ] -all
20
22 condor_vacate_job finds one or more jobs from the HTCondor job queue
23 and vacates them from the host(s) where they are currently running. The
24 jobs remain in the job queue and return to the idle state.
25 A job running under the standard universe will first produce a check‐
27 point and then the job will be killed. HTCondor will then restart the
28 job somewhere else, using the checkpoint to continue from where it left
29 off. A job running under any other universe will be sent a soft kill
30 signal (SIGTERM by default, or whatever is defined as the SoftKillSig
31 in the job ClassAd), and HTCondor will restart the job from the begin‐
32 ning somewhere else.
33 If the -fast option is used, the job(s) will be immediately killed,
35 meaning that standard universe jobs will not be allowed to checkpoint,
36 and the job will have to revert to the last checkpoint or start over
37 from the beginning.
38 If the -name option is specified, the named condor_schedd is targeted
40 for processing. If the -addr option is used, the condor_schedd at the
41 given address is targeted for processing. Otherwise, the local con‐
42 dor_schedd is targeted. The jobs to be vacated are identified by one or
43 more job identifiers, as described below. For any given job, only the
44 owner of the job or one of the queue super users (defined by the
45 QUEUE_SUPER_USERS macro) can vacate the job.
46 Using condor_vacate_job on jobs which are not currently running has no
48 effect.
49
51 -help Display usage information
52 -version
54 Display version information
55 -pool centralmanagerhostname[:portnumber]
57 Specify a pool by giving the central manager's host name and
58 an optional port number
59 -name scheddname
61 Send the command to a machine identified by scheddname
62 -addr "<a.b.c.d:port>"
64 Send the command to a machine located at "<a.b.c.d:port>"
65 cluster
67 Vacate all jobs in the specified cluster
68 cluster.process
70 Vacate the specific job in the cluster
71 user Vacate jobs belonging to specified user
73 -constraint expression
75 Vacate all jobs which match the job ClassAd expression con‐
76 straint
77 -all Vacate all the jobs in the queue
79 -fast Perform a fast vacate and hard kill the jobs
81
83 Do not confuse condor_vacate_job with condor_vacate. condor_vacate is
84 given a list of hosts to vacate, regardless of what jobs happen to be
85 running on them. Only machine owners and administrators have permission
86 to use condor_vacate to evict jobs from a given host. condor_vacate_job
87 is given a list of job to vacate, regardless of which hosts they happen
88 to be running on. Only the owner of the jobs or queue super users have
89 permission to use condor_vacate_job.
90
92 To vacate job 23.0:
93 % condor_vacate_job 23.0
95 To vacate all jobs of a user named Mary:
97 % condor_vacate_job mary
99 To vacate all standard universe jobs owned by Mary:
101 % condor_vacate_job -constraint 'JobUniverse == 1 && Owner == "mary"'
103 Note that the entire constraint, including the quotation marks, must be
105 enclosed in single quote marks for most shells.
106
108 condor_vacate_job will exit with a status value of 0 (zero) upon suc‐
109 cess, and it will exit with the value 1 (one) upon failure.
110
112 HTCondor Team
113
115 1990-2021, Center for High Throughput Computing, Computer Sciences De‐
116 partment, University of Wisconsin-Madison, Madison, WI, US. Licensed
117 under the Apache License, Version 2.0.
1188.8 Jan 26, 2021 CONDOR_VACATE_JOB(1)