1CONDOR_WATCH_Q(1) HTCondor Manual CONDOR_WATCH_Q(1)
2
6 condor_watch_q - HTCondor Manual
7 Track the status of jobs over time.
9
14 condor_watch_q [-help]
15 condor_watch_q [general options] [display options] [behavior options]
17 [tracking options]
18
20 condor_watch_q is a tool for tracking the status of jobs over time
21 without repeatedly querying the condor_schedd. It does this by reading
22 job event log files. These files may be specified directly (the -files
23 option), or indirectly via a single query to the condor_schedd when
24 condor_watch_q starts up (options like -users or -clusters).
25 condor_watch_q provides a variety of options for output formatting, in‐
27 cluding: colorized output, tabular information, progress bars, and text
28 summaries. These display options are highly-customizable via command
29 line options.
30 condor_watch_q also provides a minimal language for exiting when cer‐
32 tain conditions are met by the tracked jobs. For example, it can be
33 configured to exit when all of the tracked jobs have terminated.
34
36 If no users, cluster ids, or event logs are given, condor_watch_q will
37 default to tracking all of the current user's jobs. Thus, with no argu‐
38 ments,
39 condor_watch_q
41 will track all of your currently-active clusters.
43 To track jobs from a specific cluster, use the -clusters option, pass‐
45 ing the cluster ID:
46 condor_watch_q -clusters 12345
48 To track jobs from a specific user, use the -users option, passing the
50 user's name the actual query will be the for the Owner job ad attri‐
51 bute):
52 condor_watch_q -users jane
54 To track jobs from a specific event log file, use the -files option,
56 passing the path to the event log:
57 condor_watch_q -users /home/jane/events.log
59 To track jobs from a specific batch, use the -batches option, passing
61 the batch name:
62 condor_watch_q -batches BatchOfJobsFromTuesday
64 All of the above "tracking" options can be used together, and multiple
66 values may be passed to each one. For example, to track all of the jobs
67 that are: owned by jane or jim, in cluster 12345, or in the event log
68 /home/jill/events.log, run
69 condor_watch_q -users jane jim -clusters 12345 -files /home/jill/events.log
71 By default, condor_watch_q will never exit on its own (unless it en‐
73 counters an error or it is not tracking any jobs). You can tell it to
74 exit when certain conditions are met. For example, to exit with status
75 0 when all of the jobs it is tracking are done or with status 1 when
76 any job is held, you could run
77 condor_watch_q -exit all,done,0 -exit any,held,1
79
81 General Options
82 -help Display the help message and exit.
83 -debug Causes debugging information to be sent to stderr.
85 Tracking Options
87 These options control which jobs condor_watch_q will track, and how it
88 discovers them.
89 -users USER [USER ...]
91 Choose which users to track jobs for. All of the user's jobs
92 will be tracked. One or more user names may be passed.
93 -clusters CLUSTER_ID [CLUSTER_ID ...]
95 Which cluster IDs to track jobs for. One or more cluster ids
96 may be passed.
97 -files FILE [FILE ...]
99 Which job event log files (i.e., the log file from con‐
100 dor_submit) to track jobs from. One or more file paths may
101 be passed.
102 -batches BATCH_NAME [BATCH_NAME ...]
104 Which job batch names to track jobs for. One or more batch
105 names may be passed.
106 -collector COLLECTOR
108 Which collector to contact to find the schedd, if needed.
109 Defaults to the local collector.
110 -schedd SCHEDD
112 Which schedd to contact for queries, if needed. Defaults to
113 the local schedd.
114 Behavior Options
116 -exit GROUPER,JOB_STATUS[,EXIT_STATUS]
117 Specify conditions under which condor_watch_q should exit.
118 GROUPER is one of all, any or none. JOB_STATUS is one of ac‐
119 tive, done, idle, or held. The "active" status means "in the
120 queue", and includes jobs in the idle, running, and held
121 states. EXIT_STATUS may be any valid exit status integer.
122 To specify multiple exit conditions, pass this option multi‐
123 ple times. condor_watch_q will exit when any of the condi‐
124 tions are satisfied.
125 Display Options
127 These options control how condor_watch_q formats its output. Many of
128 them are "toggles": -x enables option "x", and -no-x disables it.
129 -groupby {batch, log, cluster}
131 How to group jobs into rows for display in the table. Must
132 be one of batch (group by job batch name), log (group by
133 event log file path), or cluster (group by cluster ID). De‐
134 faults to batch.
135 -table/-no-table
137 Enable/disable the table. Enabled by default.
138 -progress/-no-progress
140 Enable/disable the progress bar. Enabled by default.
141 -row-progress/-no-row-progress
143 Enable/disable the progress bar for each row. Enabled by de‐
144 fault.
145 -summary/-no-summary
147 Enable/disable the summary line. Enabled by default.
148 -summary-type {totals, percentages}
150 Choose what to display on the summary line, totals (the num‐
151 ber of each jobs in each state), or percentages (the percent‐
152 age of jobs in each state, of the total number of tracked
153 jobs) By default, show totals.
154 -updated-at/-no-updated-at
156 Enable/disable the "updated at" line. Enabled by default.
157 -abbreviate/-no-abbreviate
159 Enable/disable abbreviating path components to the shortest
160 somewhat-unique prefix. Disabled by default.
161 -color/-no-color
163 Enable/disable colored output. Enabled by default if con‐
164 nected to a tty. Disabled on Windows if colorama is not
165 available (https://pypi.org/project/colorama/).
166 -refresh/-no-refresh
168 Enable/disable refreshing output. If refreshing is disabled,
169 output will be appended instead. Enabled by default if con‐
170 nected to a tty.
171
173 Returns 0 when sent a SIGINT (keyboard interrupt).
174 Returns 0 if no jobs are found to track.
176 Returns 1 for fatal internal errors.
178 Can be configured via the -exit option to return any valid exit status
180 when a certain condition is met.
181
183 Center for High Throughput Computing, University of Wisconsin-Madison
184
186 HTCondor Team
187
189 1990-2023, Center for High Throughput Computing, Computer Sciences De‐
190 partment, University of Wisconsin-Madison, Madison, WI, US. Licensed
191 under the Apache License, Version 2.0.
192 Oct 02, 2023 CONDOR_WATCH_Q(1)