Thursday Exercise 4.1: Coordinating a Set of Jobs With a Simple DAG¶
The objective of this exercise is to learn the very basics of running a set of jobs, where our set is just one job.
What is DAGMan?¶
In short, DAGMan lets you submit complex sequences of jobs as long as they can be expressed as a directed acylic graph. For example, you may wish to run a large parameter sweep but before the sweep run you need to prepare your data. After the sweep runs, you need to collate the results. This might look like this, assuming you want to sweep over five parameters:
DAGMan has many abilities, such as throttling jobs, recovery from failures, and more. More information about DAGMan can be found at in the HTCondor manual.
Submitting a Simple DAG¶
For our job, we will return briefly to the sleep program.
executable = /bin/sleep arguments = 4 log = simple.log output = simple.out error = simple.error request_memory = 1GB request_disk = 1GB request_cpus = 1 queue
We are going to get a bit more sophisticated in submitting our jobs now. Let's have three windows open. In one window, you'll submit the job. In another you will watch the queue, and in the third you will watch what DAGMan does.
First we will create the most minimal DAG that can be created: a DAG with just one node. Put this into a file named
simple.dag.
Job Simple simple.sub
In your first window, submit the DAG:
username@learn $ condor_submit_dag simple.dag ----------------------------------------------------------------------- File for submitting this DAG to Condor : simple.dag.condor.sub Log of DAGMan debugging messages : simple.dag.dagman.out Log of Condor library output : simple.dag.lib.out Log of Condor library error messages : simple.dag.lib.err Log of the life of condor_dagman itself : simple.dag.dagman.log Submitting job(s). 1 job(s) submitted to cluster 61. -----------------------------------------------------------------------
In the second window, check the queue (what you see may be slightly different):
username@learn $ condor_q -nobatch -wide:80 -- Submitter: learn.chtc.wisc.edu : <128.104.100.55:9618?sock=28867_10e4_2> : learn.chtc.wisc.edu ID OWNER SUBMITTED RUN_TIME ST PRI SIZE CMD 61.0 roy 6/21 22:51 0+00:03:47 R 0 0.3 condor_dagman 62.0 roy 6/21 22:51 0+00:00:03 R 0 0.7 simple 4 10 2 jobs; 0 completed, 0 removed, 0 idle, 2 running, 0 held, 0 suspended
In the third window, watch what DAGMan does (what you see may be slightly different):
username@learn $ tail -f --lines=500 simple.dag.dagman.out 6/21/12 22:51:13 Setting maximum accepts per cycle 8. 06/21/12 22:51:13 ****************************************************** 06/21/12 22:51:13 ** condor_scheduniv_exec.61.0 (CONDOR_DAGMAN) STARTING UP 06/21/12 22:51:13 ** /usr/bin/condor_dagman 06/21/12 22:51:13 ** SubsystemInfo: name=DAGMAN type=DAGMAN(10) class=DAEMON(1) 06/21/12 22:51:13 ** Configuration: subsystem:DAGMAN local:<NONE> class:DAEMON 06/21/12 22:51:13 ** $CondorVersion: 7.7.6 Apr 16 2012 BuildID: 34175 PRE-RELEASE-UWCS $ 06/21/12 22:51:13 ** $CondorPlatform: x86_64_rhap_5.7 $ 06/21/12 22:51:13 ** PID = 5812 06/21/12 22:51:13 ** Log last touched 6/21 22:51:00 06/21/12 22:51:13 ****************************************************** 06/21/12 22:51:13 Using config source: /etc/condor/condor_config 06/21/12 22:51:13 Using local config sources: 06/21/12 22:51:13 /etc/condor/config.d/00-chtc-global.conf 06/21/12 22:51:13 /etc/condor/config.d/01-chtc-submit.conf 06/21/12 22:51:13 /etc/condor/config.d/02-chtc-flocking.conf 06/21/12 22:51:13 /etc/condor/config.d/03-chtc-jobrouter.conf 06/21/12 22:51:13 /etc/condor/config.d/04-chtc-blacklist.conf 06/21/12 22:51:13 /etc/condor/config.d/99-osg-ss-group.conf 06/21/12 22:51:13 /etc/condor/config.d/99-roy-extras.conf 06/21/12 22:51:13 /etc/condor/condor_config.local 06/21/12 22:51:13 DaemonCore: command socket at <128.104.100.55:60417> 06/21/12 22:51:13 DaemonCore: private command socket at <128.104.100.55:60417> 06/21/12 22:51:13 Setting maximum accepts per cycle 8. 06/21/12 22:51:13 DAGMAN_USE_STRICT setting: 0 06/21/12 22:51:13 DAGMAN_VERBOSITY setting: 3 06/21/12 22:51:13 DAGMAN_DEBUG_CACHE_SIZE setting: 5242880 06/21/12 22:51:13 DAGMAN_DEBUG_CACHE_ENABLE setting: False 06/21/12 22:51:13 DAGMAN_SUBMIT_DELAY setting: 0 06/21/12 22:51:13 DAGMAN_MAX_SUBMIT_ATTEMPTS setting: 6 06/21/12 22:51:13 DAGMAN_STARTUP_CYCLE_DETECT setting: False 06/21/12 22:51:13 DAGMAN_MAX_SUBMITS_PER_INTERVAL setting: 5 06/21/12 22:51:13 DAGMAN_USER_LOG_SCAN_INTERVAL setting: 5 06/21/12 22:51:13 allow_events (DAGMAN_IGNORE_DUPLICATE_JOB_EXECUTION, DAGMAN_ALLOW_EVENTS) setting: 114 06/21/12 22:51:13 DAGMAN_RETRY_SUBMIT_FIRST setting: True 06/21/12 22:51:13 DAGMAN_RETRY_NODE_FIRST setting: False 06/21/12 22:51:13 DAGMAN_MAX_JOBS_IDLE setting: 0 06/21/12 22:51:13 DAGMAN_MAX_JOBS_SUBMITTED setting: 0 06/21/12 22:51:15 DAGMAN_MAX_PRE_SCRIPTS setting: 0 06/21/12 22:51:15 DAGMAN_MAX_POST_SCRIPTS setting: 0 06/21/12 22:51:15 DAGMAN_ALLOW_LOG_ERROR setting: False 06/21/12 22:51:15 DAGMAN_MUNGE_NODE_NAMES setting: True 06/21/12 22:51:15 DAGMAN_PROHIBIT_MULTI_JOBS setting: False 06/21/12 22:51:15 DAGMAN_SUBMIT_DEPTH_FIRST setting: False 06/21/12 22:51:15 DAGMAN_ALWAYS_RUN_POST setting: True 06/21/12 22:51:15 DAGMAN_ABORT_DUPLICATES setting: True 06/21/12 22:51:15 DAGMAN_ABORT_ON_SCARY_SUBMIT setting: True 06/21/12 22:51:15 DAGMAN_PENDING_REPORT_INTERVAL setting: 600 06/21/12 22:51:15 DAGMAN_AUTO_RESCUE setting: True 06/21/12 22:51:15 DAGMAN_MAX_RESCUE_NUM setting: 100 06/21/12 22:51:15 DAGMAN_WRITE_PARTIAL_RESCUE setting: True 06/21/12 22:51:15 DAGMAN_DEFAULT_NODE_LOG setting: null 06/21/12 22:51:15 DAGMAN_GENERATE_SUBDAG_SUBMITS setting: True 06/21/12 22:51:15 ALL_DEBUG setting: 06/21/12 22:51:15 DAGMAN_DEBUG setting: 06/21/12 22:51:15 argv[0] == "condor_scheduniv_exec.61.0" 06/21/12 22:51:15 argv[1] == "-Lockfile" 06/21/12 22:51:15 argv[2] == "simple.dag.lock" 06/21/12 22:51:15 argv[3] == "-AutoRescue" 06/21/12 22:51:15 argv[4] == "1" 06/21/12 22:51:15 argv[5] == "-DoRescueFrom" 06/21/12 22:51:15 argv[6] == "0" 06/21/12 22:51:15 argv[7] == "-Dag" 06/21/12 22:51:15 argv[8] == "simple.dag" 06/21/12 22:51:15 argv[9] == "-CsdVersion" 06/21/12 22:51:15 argv[10] == "$CondorVersion: 7.7.6 Apr 16 2012 BuildID: 34175 PRE-RELEASE-UWCS $" 06/21/12 22:51:15 argv[11] == "-Force" 06/21/12 22:51:15 argv[12] == "-Dagman" 06/21/12 22:51:15 argv[13] == "/usr/bin/condor_dagman" 06/21/12 22:51:15 Default node log file is: </home/roy/condor/simple.dag.nodes.log> 06/21/12 22:51:15 DAG Lockfile will be written to simple.dag.lock 06/21/12 22:51:15 DAG Input file is simple.dag 06/21/12 22:51:15 Parsing 1 dagfiles 06/21/12 22:51:15 Parsing simple.dag ... 06/21/12 22:51:15 Dag contains 1 total jobs 06/21/12 22:51:15 Sleeping for 12 seconds to ensure ProcessId uniqueness 06/21/12 22:51:27 Bootstrapping... 06/21/12 22:51:27 Number of pre-completed nodes: 0 06/21/12 22:51:27 Registering condor_event_timer... 06/21/12 22:51:28 Sleeping for one second for log file consistency 06/21/12 22:51:29 MultiLogFiles: truncating log file /home/roy/condor/simple.log 06/21/12 22:51:29 Submitting Condor Node Simple job(s)...
Here's where the job is submitted
06/21/12 22:51:29 submitting: condor_submit
-a dag_node_name' '=' 'Simple
-a +DAGManJobId' '=' '61
-a DAGManJobId' '=' '61
-a submit_event_notes' '=' 'DAG' 'Node:' 'Simple
-a DAG_STATUS' '=' '0
-a FAILED_COUNT' '=' '0
-a +DAGParentNodeNames' '=' '"" submit
06/21/12 22:51:30 From submit: Submitting job(s).
06/21/12 22:51:30 From submit: 1 job(s) submitted to cluster 62.
06/21/12 22:51:30 assigned Condor ID (62.0.0)
06/21/12 22:51:30 Just submitted 1 job this cycle...
06/21/12 22:51:30 Currently monitoring 1 Condor log file(s)
06/21/12 22:51:30 Event: ULOG_SUBMIT for Condor Node Simple (62.0.0)
06/21/12 22:51:30 Number of idle job procs: 1
06/21/12 22:51:30 Of 1 nodes total:
06/21/12 22:51:30 Done Pre Queued Post Ready Un-Ready Failed
06/21/12 22:51:30 === === === === === === ===
06/21/12 22:51:30 0 0 1 0 0 0 0
06/21/12 22:51:30 0 job proc(s) currently held
06/21/12 22:55:05 Currently monitoring 1 Condor log file(s)
Here's where DAGMan noticed that the job is running
06/21/12 22:55:05 Event: ULOG_EXECUTE for Condor Node Simple (62.0.0) 06/21/12 22:55:05 Number of idle job procs: 0 06/21/12 22:55:10 Currently monitoring 1 Condor log file(s) 06/21/12 22:55:10 Event: ULOG_IMAGE_SIZE for Condor Node Simple (62.0.0) 06/21/12 22:56:05 Currently monitoring 1 Condor log file(s) 06/21/12 22:56:05 Event: ULOG_IMAGE_SIZE for Condor Node Simple (62.0.0)
Here's where DAGMan noticed that the job finished.
06/21/12 22:56:05 Event: ULOG_JOB_TERMINATED for Condor Node Simple (62.0.0) 06/21/12 22:56:05 Node Simple job proc (62.0.0) completed successfully. 06/21/12 22:56:05 Node Simple job completed 06/21/12 22:56:05 Number of idle job procs: 0 06/21/12 22:56:05 Of 1 nodes total: 06/21/12 22:56:05 Done Pre Queued Post Ready Un-Ready Failed 06/21/12 22:56:05 === === === === === === === 06/21/12 22:56:05 1 0 0 0 0 0 0 06/21/12 22:56:05 0 job proc(s) currently held
Here's where DAGMan noticed that all the work is done.
06/21/12 22:56:05 All jobs Completed! 06/21/12 22:56:05 Note: 0 total job deferrals because of -MaxJobs limit (0) 06/21/12 22:56:05 Note: 0 total job deferrals because of -MaxIdle limit (0) 06/21/12 22:56:05 Note: 0 total job deferrals because of node category throttles 06/21/12 22:56:05 Note: 0 total PRE script deferrals because of -MaxPre limit (0) 06/21/12 22:56:05 Note: 0 total POST script deferrals because of -MaxPost limit (0) 06/21/12 22:56:05 **** condor_scheduniv_exec.61.0 (condor_DAGMAN) pid 5812 EXITING WITH STATUS 0
Now verify your results:
username@learn $ cat simple.log 000 (062.000.000) 06/21 22:51:30 Job submitted from host: <128.104.100.55:9618?sock=28867_10e4_2> DAG Node: Simple ... 001 (062.000.000) 06/21 22:55:00 Job executing on host: <128.104.58.36:46761> ... 006 (062.000.000) 06/21 22:55:09 Image size of job updated: 750 3 - MemoryUsage of job (MB) 2324 - ResidentSetSize of job (KB) ... 006 (062.000.000) 06/21 22:56:00 Image size of job updated: 780 3 - MemoryUsage of job (MB) 2324 - ResidentSetSize of job (KB) ... 005 (062.000.000) 06/21 22:56:00 Job terminated. (1) Normal termination (return value 0) Usr 0 00:00:00, Sys 0 00:00:00 - Run Remote Usage Usr 0 00:00:00, Sys 0 00:00:00 - Run Local Usage Usr 0 00:00:00, Sys 0 00:00:00 - Total Remote Usage Usr 0 00:00:00, Sys 0 00:00:00 - Total Local Usage 57 - Run Bytes Sent By Job 608490 - Run Bytes Received By Job 57 - Total Bytes Sent By Job 608490 - Total Bytes Received By Job Partitionable Resources : Usage Request Cpus : 1 Disk (KB) : 750 750 Memory (MB) : 3 3 ...
Looking at DAGMan's various files, we see that DAGMan itself ran as a job (specifically, a "scheduler" universe job).
username@learn $ ls simple.dag.* simple.dag.condor.sub simple.dag.dagman.log simple.dag.dagman.out simple.dag.lib.err simple.dag.lib.out username@learn $ cat simple.dag.condor.sub # Filename: simple.dag.condor.sub # Generated by condor_submit_dag simple.dag universe = scheduler executable = /usr/bin/condor_dagman getenv = True output = simple.dag.lib.out error = simple.dag.lib.err log = simple.dag.dagman.log remove_kill_sig = SIGUSR1 +OtherJobRemoveRequirements = "DAGManJobId == $(cluster)" # Note: default on_exit_remove expression: # ( ExitSignal =?= 11 || (ExitCode =!= UNDEFINED && ExitCode >=0 && ExitCode <= 2)) # attempts to ensure that DAGMan is automatically # requeued by the schedd if it exits abnormally or # is killed (e.g., during a reboot). on_exit_remove = ( ExitSignal =?= 11 || (ExitCode =!= UNDEFINED && ExitCode >=0 && ExitCode <= 2)) copy_to_spool = False arguments = "-f -l . -Lockfile simple.dag.lock -AutoRescue 1 -DoRescueFrom 0 -Dag simple.dag -CsdVersion $CondorVersion:' '7.7.6' 'Apr' '16' '2012' 'BuildID:' '34175' 'PRE-RELEASE-UWCS' '$ -Force -Dagman /usr/bin/condor_dagman" environment = _CONDOR_DAGMAN_LOG=simple.dag.dagman.out;_CONDOR_MAX_DAGMAN_LOG=0 queue
If you want to clean up some of these files (you may not want to, at least not yet), run:
username@learn $ rm simple.dag.*
Challenge¶
- What is the scheduler universe? Why does DAGMan use it?