1condor_drain(1)             General Commands Manual            condor_drain(1)
2
3
4

Name

6       condor_drainControl draining of an execute machine
7

Synopsis

9       condor_drain[-help]
10
11       condor_drain[-debug]  [-pool  pool-name]  [-graceful  | -quick | -fast]
12       [-resume-on-completion] [-check expr] [-start expr] machine-name
13
14       condor_drain[-debug] [-pool pool-name] -cancel[-request-id id] machine-
15       name
16

Description

18       condor_drainis  an  administrative command used to control the draining
19       of all slots on an execute machine. When a machine is draining, it will
20       not  accept  any  new jobs unless the -startexpression specifies other‐
21       wise. Which machine to drain is specified by the argument machine-name,
22       and will be the same as the machine ClassAd attribute Machine.
23
24       How currently running jobs are treated depends on the draining schedule
25       that is chosen with a command-line option:
26
27       -graceful
28
29          Initiate a graceful eviction of the job.  This  means  all  promises
30          that  have been made to the job are honored, including MaxJobRetire‐
31          mentTime. The eviction of jobs is coordinated to reduce  idle  time.
32          This  means  that  if one slot has a job with a long retirement time
33          and the other slots have jobs with  shorter  retirement  times,  the
34          effective  retirement time for all of the jobs is the longer one. If
35          no draining schedule is specified, -gracefulis chosen by default.
36
37
38
39
40
41       -quick
42
43          MaxJobRetirementTimeis not honored. Eviction of jobs is  immediately
44          initiated. Jobs are given time to shut down and produce checkpoints,
45          according to the usual policy, that is, given  by  MachineMaxVacate‐
46          Time.
47
48
49
50
51
52       -fast
53
54          Jobs  are immediately hard-killed, with no chance to gracefully shut
55          down or produce a checkpoint.
56
57
58
59
60
61       If you specify -graceful, you may also specify -start. On a gracefully-
62       draining  machine,  some  jobs  may  finish  retiring before others. By
63       default, the resources used by the newly-retired  jobs  do  not  become
64       available  for  use  by other jobs until the machine exits the draining
65       state (see below). The -startexpression you supply replaces the  drain‐
66       ing  machine's  normal STARTexpression for the duration of the draining
67       state, potentially making those resources available.  See  section  for
68       more information.
69
70       Once  draining  is  complete,  the  machine will enter the Drained/Idle
71       state. To resume normal operation (negotiation) at  that  time  or  any
72       previous  time  during  draining,  the  -canceloption  may be used. The
73       -resume-on-completionoption results in automatic resumption  of  normal
74       operation  once draining has completed, and may be used when initiating
75       draining. This is useful for forcing a  machine  with  a  partitionable
76       slots  to  join  all  of  the resources back together into one machine,
77       facilitating de-fragmentation and whole machine negotiation.
78

Options

80       -help
81
82          Display brief usage information and exit.
83
84
85
86       -debug
87
88          Causes debugging information to be sent  to  stderr,  based  on  the
89          value of the configuration variable TOOL_DEBUG.
90
91
92
93       -pool pool-name
94
95          Specify  an  alternate  HTCondor  pool,  if  the  default one is not
96          desired.
97
98
99
100       -graceful
101
102          (the default) Honor the maximum vacate and retirement time policy.
103
104
105
106       -quick
107
108          Honor the maximum vacate time, but not the retirement time policy.
109
110
111
112       -fast
113
114          Honor neither the maximum vacate time policy nor the retirement time
115          policy.
116
117
118
119       -resume-on-completion
120
121          When  done  draining, resume normal operation, such that potentially
122          the whole machine could be claimed.
123
124
125
126       -check expr
127
128          Abort draining, if expris not true for all slots to be drained.
129
130
131
132       -start expr
133
134          The STARTexpression to use while the machine is draining. You  can't
135          reference the machine's existing STARTexpression.
136
137
138
139       -cancel
140
141          Cancel  a  prior draining request, to permit the condor_negotiatorto
142          use the machine again.
143
144
145
146       -request-id id
147
148          Specify a specific draining request to cancel, where idis  given  by
149          the DrainingRequestIdmachine ClassAd attribute.
150
151
152

Exit Status

154       condor_drainwill exit with a non-zero status value if it fails and zero
155       status if it succeeds.
156

Author

158       Center for High Throughput Computing, University of Wisconsin-Madison
159

161       Copyright (C) 1990-2019 Center for High Throughput Computing,  Computer
162       Sciences  Department, University of Wisconsin-Madison, Madison, WI. All
163       Rights Reserved. Licensed under the Apache License, Version 2.0.
164
165
166
167                                     date                      condor_drain(1)
Impressum