Openpbs Versions Save

You can download the packages from https://www.openpbs.org/Download.aspx#download

Builds are available for:

OpenSuse Leap 15.4 RockyLinux 8.8 Ubuntu 18.04 Ubuntu 20.04

be361a18 (tag: v23.06.06) update openpbs version (#2591) 9efa7724 Ubuntu 22.04 Jammy Support (#2583) 1fa2455c Fix server crash on delete array job with duplicate sub job ids (#2581) f899d6c6 Fixed Compilation Warnings (#2577) 8b640850 fix scheduler registering with kerberos (#2574) bbeb7ba4 fix server crash due to invalid python writes (#2565) c2501d94 Introduce postqueuejob server hook (#2567) 38a623f6 Fix: run job which got -default- group (#2558) a9c24e41 reject job if uid is not found instead of offlining vnode (#2557) df4c3206 Fixing upgrades (#2551) b4b12543 Management Debug Hook Fix (#2550) 5826dd5a Custom Installed Python - 3.9 (#2549) 8e4201cd Fix check for running systemd (#2548) af2b891f fix typo listong -> listing (#2543) 8f05f1bf Fix build on macOS (#2544) d54642c7 Avoid hardcoding default configuration variables (#2547) 5882a599 Fix su invocations to work on both linux and bsd (#2546) 4e5bd9f3 Add includes to mom to fix dependencies (#2545) f8e501ad Install PBS env shell scripts in PBS_EXEC/etc (#2537)

OpenPBS v22.05.11 changes

You can download the packages from https://www.openpbs.org/Download.aspx#download 32

Builds are available for

OpenSuse Leap 15.3
RockyLinux 8.5
Ubuntu 18.04
Ubuntu 20.04

You will notice that this is a new versioning system. We have moved to a date-based versioning system for OpenPBS builds. This makes it easier for users to build their own versions while maintaining consistent versioning for their build.

There have been over 500 changes to Open PBS since the last release: Commits · openpbs/openpbs · GitHub 16

Some of the highlights include

New hook events
    Management event - Any set/unset by qmgr
    Modify vnode event
    Job obit event
    Reservation Begin event
    Reservation confirm event
    Modify Reservation event
Support for Nvidia A100 GPUs
Added support for Rocky Linux and Ubuntu 20
Bug fixes
Refactoring of the code
Fixed memory leaks
New automated tests

You will notice some changes to the site too over the next few months and we update the online material. Thanks again to all the members of the community who have contributed to that or make suggestions.

OpenPBS v20.0.1 changes

48bcb66f Fix for PBS_SUPPORTED_AUTH_METHODS not honored (#1939) 4e0f9b89 Fix for pgsql version parsing in pbs_habitat (#1938)

OpenPBS v20.0 Highlights

Important upgrade note: rpm -U cannot be used to upgrade from an earlier release (because the software name has changed to OpenPBS). Overlay and migration upgrades from v19.1 to v20.0 should work. To do an overlay upgrade: (1) stop PBS, (2) make a backup of your installation, (3) uninstall the older version of PBS via “rpm -e”, (4) install the new OpenPBS via “rpm -i”, and (5) start PBS up again.

What’s New?

• Faster scheduling via multithreading, better job dispatch (~30x faster), improvements preempting large numbers of jobs (~2x faster), and asynchronous job updates.
• Faster/better turnaround and utilization via a new BETA feature: “runone” dependencies, which runs the first job possible among a set of alternatives, i.e., run job X or Y or Z, whichever will start first.
• Faster, more scalable, hook throughput via pre-cached job data and a non-blocking execjob_end event; plus, additional hook events, including pre-resume and post-suspend
• Native PBS support for the Cray Shasta platform
• Plugin framework (hooks) upgraded to support Python 3
• Numerous advance reservations enhancements: supporting multisched, freeing idle reservations via timeouts, replacing failed nodes in running reservations, changing durations, updating allowed users/groups, explicitly freeing some allocated resources (to shrink a reservation), and creating maintenance reservations regardless of node state
• Long-waiting jobs can “keep their place in the queue” despite startup faults or human errors by converting these jobs to new “start” and “now” reservations (which an admin can do automatically, e.g., based on job size)
• Accounting records now include full submission information (including dependencies), updates on qalter, and suspend/resume event information
• Enhanced cgroup functionality: flexible configuration for heterogeneous clusters, control of memory.swappiness, support for “cpu” subsystem, and replacement of old cpuset MOM on HPE systems.
• Plus, preemption by deletion for non-rerunnable jobs, additional qstat -w (wide) options, smaller MOM disk footprint, ability to hold subjobs, consistent log_events attribute for the scheduler, select statement syntax for pbs_release_nodes, support for multiple authentication methods, and more.

View the full list of enhancements and bug fixes in the commit log.

PBS Pro v19.1.3 Release

Commits in this release: 33345bea Change pbs version from 19.1.2 to 19.1.3 (#1346) 80f33a47 Make mom accept only privileged requests (#1340) 0f0fa0ac Build fails in opensuse leap-15 due to third party issues (#1331) (#1342)

** Performing an overlay upgrade between 19.1.1 and 19.1.2 will result in the pbs_mom losing track of running job processes, causing the jobs to appear to run forever to PBS Pro. You must not upgrade from 19.1.1 to 19.1.2 while any PBS jobs are running. Upgrades from versions prior to 19.1.1 do not have this problem, and it is therefore safe to allow single host jobs to remain running during an overlay upgrade provided the published instructions are followed. Upgrades from 19.1.2 onward will not be impacted by this.

Commits in this release: f1e05f56 PBS doesn't finish already running jobs after overlay upgrade (#1049) (#1154) f0de9fc4 database shutdown / server abort relating to duplicate queue creation (#1151) 53382655 Sched crash when array job and reservation is submitted (#1158)

PBS Pro v18.1.4 Release

Commits in this release: 4ddbf807 PBS Installation fails on Open SUSE Leap 15.0 with libical dependency (#1054) 13cfd53d Update PBS Pro version number to 18.1.4 ba1c10b9 Bugfix: reading version correctly from multiline output of psql -V (#1045) e30ad97d Partial cherry pick of PR 759 for ical3

Changes in this release (from 18.1.3 to v19.1.1)

c69bd91 PP-1326: fix dirname: missing operand error printed from PTL related init scripts 2ca43d8 update test case to avoid race condition a0e5e44 To get the recommended unit file directory to place the pbs.service file for systemctl command 5f60875 Fixing test failures post manager() interface change dd79ea1 PBS cgroups hook gives error - 'dict' object is not callable 205d9be Remove day argument from log_match efc7e95 Modifying test_qrun_hangs to address race condition 34f9830 PTL reverts multi-server config incorrectly cac732b Refactoring: moving find_string to libutil 726499e qmove fix for migration upgrade 9c96636 TestMultipleSchedulers.test_fairshare_usage fails due to race condition ca53e30 test_subjob_comments failing due to race condition 3009492 Upgrade fails from 18.1.3 to 19.1.1 with postgresql error on Open SUSE 54f497e Added new performance test for node buckets b4016eb libjob support for SLES15 f46e4b3 Making logging to asynchronous for the daemons 304b199 Race conditions in PTL tests when using manager() 51bb897 Updates for adding a sample hook for Reliable Job Startup 714ae76 Fix crash in tpp_transport_terminate() after pthread_at_fork() executes 4e9189f Fix crash when one daemon with compression talks to another without compression 287f88e increase the default timeout for PTL tests 145fee0 automate pbsnodes -a test 23cbf3a Fix preemption order in tests 34d5c6c Fix test_suspend_time_not_counted_in_walltime by removing hardcoded walltime values 0884b50 PBS Installation fails on Open SUSE Leap 15.0 with libical dependency 2950f33 tests were failing as -h was passed with an option 8080052 SLES15-Cray gcc warning fix 76fc34f TestMultipleSchedulers.test_equiv_suspend_jobs failing due to race condition 8b972d2 fixing the docstring for one of the test in pbs_offline_vnodes.py eb29906 Fix scheduler not logging reason denying reservation ba6e7bd pbs_equiv_classes.py tests has too few max_attempts for log_match 4edf7c0 PTL expect doesn't work if provided the wrong case 1f8828f PBS schema upgrade from 1.3 to 1.4 299bcfe Interleave TCP/TPP better f8e9b76 job submission was failing due to incorrect format 16d36f1 Database refactoring to eliminate server name and server hostname 1b3a4fe Add test for subjobs in wrong substate 31208f7 few tests were failing since system has 2 paths for cpuset dcbd814 Allowing . in job name 18c1fbc Updates for PTL package c4490e2 Refactoring: removing pbs_config.h includes from .h files 941bbb5 Fix for qrun hang 3f6f235 Disable Nagles Algorithm in scheduler 25deae4 Updating INSTALL readme file to have postgresql-contrib 5412674 Use short hostname in pbs.conf while reverting pbs.conf changes 1133462 job_name validation and PTL test fef8bf0 Handle rpmlint warning: no-changelogname-tag 3414c3a test_queue_deletion_after_terminated_subjob fails if server restart is slow 60d77bb Package for ptl installation a87f792 TestFairshare.test_pbsfs fails intermittently ac5333e copyright header was not accurate for some files 21a1168 Missing call to TestFunctional.setup() d0f0169 test_prologue_hook_does_not_execute_twic is failing due to incorrect mom 1ebf32d Enhance pbs_snapshot to capture qstat & pbsnodes json outputs e270920 Database refactoring, creation and using API for the same c47231d create_import_hook does not check if hook files are copied to moms 0e1a1f4 PTL doesn't revert pbs.conf in setUp b15ce75 json report of PTL test run 9e1a96f Updates to tests for reliable job startup feature on cgroups dbb2e7f Fix test test_last_pset_can_never_run a2af430 Only update jobs that MoM is aware of a62a350 Logmatch error due to hostname usage rather than shortname e6b4b19 Handle rpmlint: non-conffile-in-etc,non-executable-script,explicit-lib-dependency,dir-or-file-in-opt fe355a8 Entity limits not working for job arrays bd75e3a Fix qsub with multiple arguments 71a1311 asynchronous mom job saves 5e4c2c3 handle windows style path and use snprintf for commit 4ae25fd 24c10cb TestMultipleSchedulers.test_backfill_per_scheduler fails intermittently due to race condition 7a8e0f0 fix server memory leak in fix_indirectness involving indirect resources 03d5e88 missing argument to the function in pbs_cgroups_stress.py test 43e5832 few tests from pbs_cgroups_hook.py were failing due to short job sleep time 4fc08f9 Create job purge as different process 687e7c5 Fix and PTL testcase for rstat long resv display 4ae25fd fixes to job array wrt -o,-e,-R and direct write qsub options 47093b4 Calling WSAStartup to initiate use of Winsock + adding batch script to differentiate server and host 7879b0e Tests of TestQrun is failing while running qrun command e3023ed TPP's use of MT-unsafe function netaddr() causes intermittent failures in name resolution f0b8f36 Address cpuset platform gcc warnings 692ead3 FQDN/TPP address cache causes problems with dynamic hostname/ip environments 76ca6e2 Fix race condition in test_job_array 30877dd test_emails() fails to read email file 3e565ed TestNodePartition fails due to incorrect node id 0197d9b Fix and PTL testcase for missing node attr 374dd7a Fix rpmlint warning: non-standard-executable-perm/setuid-binary c040ac2 update copyright header 26032c6 Jobid/Resvid wrap it to zero after server restart multiple times consecutively ed47e9f OSS PBS failed to start daemons on postgres upgrade externally d57ed81 Fixing timed out issues in TestNodeBuckets by deleting sched log in setup aa30345 Adding tests for server periodic hook 91ecddc Tests for reliable job startup on Cray and cpusets 88ffaca Job may get marked as 'Never run' c55d126 Some tests in TestQstatFormats do not get qstat attributes on the fly 3ac7ae3 Scheduler requeue/checkpoint fails when RRTROS is set eaa0998 PBS server reports expired subjobs to scheduler 7119b69 Add a M4 macro to enable online data compression in TPP 5263ad2 Set PBS_HOOK_CONFIG_FILE correctly in MoM 9e805f0 PP-928: Reliable Job Startup 4de36b6 Fixed race condition in node buckets test test_psets f62d2fe test_cgroups_race_condition do not skip if cpuset or memory cgroups are not present 430a019 dead code cleanup - dce/kerberos and gridproxy/gss 5f53c75 TestJobArray.test_subjob_comments_with_history is failing due to race condition ce0cb14 PP-289: Unique Jobid's upto 1 trillion 07b3d14 Allow extra user plugins in pbs_benchpress 28be0cf Race condition in log_match f429dd1 Fix non-unique session id on Cray on ARM node 1c71ce0 Fix invalid read of size 1 under parse_at_list():lib/Libcmds/parse_at.c b2a0f10 Log match failures in TestPbsHookSetJobEnv d1fd159 Race condition in TestEligibleTime.test_job_array 13af16d memory debug logs 828485e Add new OSes in CI and remove all old OSes and hacks from Travis 12de2af fix json escape quotes 762e3da Improving preemption performance 1ec09dd HPMC support for power management 94444ae Address errors from gcc 8 with -Wall -Werror 5408e47 Sched skips marking node for provisioning 6211604 Combine related qstat tests in a single test suite 92bfd6d Tests in Test_acl_host_moms fail when acl_hosts is not FQDN a0fc038 PP-1310: Tests of TestSchedSubjobBadstate is failing on cpuset platform due to race conditions 2915208 Updates pbs_holidays file for 2018 48a8276 Enhance pbs_snapshot to capture remote data for --with-sudo option 3f1824a Fix for mom crash 0e3b6cb PP-1308: 'test_hold_time_not_counted_in_walltime' of TestMomWalltime is failing intermittently while checking the walltime of a job 54cd559 Update copyright header on PbsPro files 3677929 Fixing node bucket performance test and added checkpoint test b7beac9 Resolve req_resvSub() memory leak 14ceb11 memory is not freed when action function fails in mgr_set_attr 924f2fd PP-1307: Tests TestSisterMom Fails as it wasn't able to find pbsdsh path c0dbae7 sched crashes while reordering nodes 4ab92ab Race condition in TestJobComment 390f05e TestAcctlogRescUsedWithTwoMomHooks not skipped when 2 moms are not provided 2c5fefc Timing issues in TestSoftWalltime de10f53 Windows Fix: default_qsub_arguments -V 185074f qsub fails with buffer overflow error 9d06112 PP-1305: A race in suspend/resume at end of job succeed in suspending but leave job in limbo 1aebaf6 buckets failed to match with placement sets and nodes associated with queues 80b9924 Fix unknown resource memory leak fa15e9e misuse of qrun -H crashes server b55857a Copyright updated for files 373f704 Adding new option --with-sudo to pbs_snpshot 66bd8c1 Added ASAP reservation tests 9b7eb46 autogen fails on RHEL-6 c10a91c Provide a pkg-config file libpbs. 7847e71 default_qsub_arguments -V not working 38be635 Fairshare test timing out 05347b3 First phase of code cleanup, removing dead code for unsupported platforms 01850f4 TestOfflineVnode - Race condition in momhook file copy 587d300 pbs_mom not starting; set_enforcement fails to handle arguments without value 5b3c404 analyze_server_log does not return any values if only start time is specified 3cfd366 LIBICAL_API2 macro is defined when libical version 1 is passed for build 7478672 PP-1026: Stranded array subjobs (commit 2: adding more logs) ce1d366 A qstat job_id.serverA@serverB does not work as specified d27ef9c pbs.conf variables could not be overridden by environment variables for pbs_server 9b41a8e Scheduler may oversubscribe node while confirming reservation ad6b982 TestCgroupsHook are still failing if cpuset cgroup is there but no memory 8efab44 pbs_snapshot fails anonymizing accounting records c8ed25e Build PBS Pro on OpenSUSE Tumbleweed 0be1f0c qstat slowdown due to excessive memory allocation (and zeroing) when displaying jobs using qsub -- /command submission style 2ad0a0a Added microsec logging test for env variable 3b4fc44 PP-1026: Stranded array subjobs after communication hiccup eb4f302 PP-992, PP-1296: Function add_to_resource_group is adding unwanted characters and doesn't HUP the scheduler internally 41d41d1 Add simple test in AppVeyor 6ebc2c3 Fix for server periodic hook event 1ef25f4 fix pbsprobe buffer overflow b8eb6ff Added test for run_count<20 eccaa47 pbs_snapshot fails when run from a hook 882f33a fix qselect -q buffer overflow 3bba2d2 Failure of any server hook event that is combined with server periodic hook event 69c04a4 checkpoint_abort not working when server was restarted using qterm immediate and delay. 7318eb3 Sometimes test case SmokeTest.test_fairshare_enhanced fails due to race condition 6904950 Tests for pbsnodes 29d2185 PP-1291: Enhance test pbs_basil_support.py to handle error in retklist() and teardown() 92314e0 Test "test_mom_down" and "test_reservation" fail while verifying job attributes 275698d Added extra test for preempt with checkpoint 88fbdfc Fix the typos in the description of pbs_hook_debug_nocrash 8769172 TestCgroupsHook not skipping if no cgroups mounted 9dd6854 Add missing space in variable name f2277fe test_cgroup_cpuset is failing intermittently on certain platform 5544b33 Fixed bugs in TestPbsNodeRampDown involving shell execution, log matching. 7946441 Skip running Test_systemd on unsupported platforms 8f14821 PP-1292: On Windows, pbsdsh fails with 'pbsdsh: tm_init failed, rc = not connected (17002) (17002)' 1572fa9 PP-1287: do not purge moved job from history before the job is finished d7cd5c4 cpuset: When cpuset version >=4 make sure libbitmask is available and used bfa66a5 ical: Replace deprecated icaltime_from_timet 173e312 Fix test_passing_shell_function() to fully test the passed function. c589e6d PP-1254: PBS starts in the background even when started using init script much after the system startup f040dd4 server retries to run a job more than 20 times 55e7ae2 'pbsnodes -o' offlines ALL vnodes/compute nodes on a Cray X* series 2ce1aee fix qsub -v mess up shell function formatting bug 2680c86 Additional PTL tests for checkpoint-abort. 63e8021 Fix qsub failure on Windows due to return value of _snprintf() being different from snprintf() b71e422 fix pbs_init_script test to avoid sudo issues and add tearDown to reset PBS default environment eef305f test_cgroup_cpuset failing with hyperthreads present 93ecf66 Refactor qsub.c 9bebc14 Fixed scheduler crash with ASAP reservations fa13fba checkpoint_abort not working 0f530fa Increase timeout for test_cgroups_race_condition c4d3ad7 Test to check memory.use_hierarchy is only set to 1 64213e8 Scheduler reports preempted job as Can Never Run for one cycle a9026ae updated test_preemption further e4454ac Bug fix for Scheduler does a reconfigure after restart of pbs services causing delayed scheduling cycle b15df97 fix clang warning: use of logical OR with constant if statement always true 4211656 node bucket host sets crash cc77449 Fix log_match failures in slower testbed machines by leaving max_attempts to default value 5ee7f11 Adding backward compatibility to pbs_snapshot ed1ecb7 Node stays in state=free with excl resv on server restart d63d892 PP-1278: Need a skip decorator in PTL which will skip the test without any condition. d90437c If job in resv exceeds resv end time, future resv fails to confirm 293bce8 Man page updates 237c6ba Node Buckets: A new node allocation algorithm for excl jobs 1410a36 Additional tests for power on off and power ramp rate f035b9a Check if hook is allowed to change a vnode before trying to cross-link a vnode with MoM fef5595 duplicate id test fixed 484577a update test_preemption of pbs_equiv_class to avoid the race condition a1e4336 Added test where hook is failing due to exception. 877f89e qsub: parameter -v with commas in value is broken 47de87c Include pbs_sleep for linux distribution 79ba2fe Added code to verify that snapshot captures the schedulers at non-default location fc24927 changes to pbs_cgroups_hook test cases for CRAY 149c642 Server state going to "Scheduling" in Failover setup 3a1775c near term excl resvs interfere with longer term resvs c1bfa42 Adding generic library function - break_delimited_str 2db700a PP-913 Reservation End hook event 9cc3776 PP-1236: few test from 'TestPbsResvAlter' failed due to Bad time specifications error 76ad82b Fix some cgroups tests ba27370 prevent using GPUs without user awareness c5223c1 small changes in cgroups tests 97108d4 Additional tests for Job arrays 231e6d2 Refactor cgroup tests and fix node creation 5d209c6 Used create_temp_file to create temp file e4dd4c1 Add the -b option to the PTL test's aprun invocation fcfda11 PP-261: Microsecond Logging c076570 Fix mom crash when deleting job in exiting state. 82ffc39 Fix gpu discovery and assignment for cgroups hook 3cb2bb2 PP-993: When PTL starts or restarts mom, after tests, the mom left behind fails to execute mom hooks da94148 remove extra destroying of a lock in tpp 91a392a test_cgroup_periodic_update failing in some cases 9ebc93d PP-917: PTL server.create_vnodes() with delall=False not validating correctly 637ba43 fix for mom crash when ALPS returns invalid XML de321b7 PTL test added to check fairshare usage in TestMultipleSchedulers c33afa7 PP-973: PTL is not reverting the managers and operators in setUp b177e5c Additional tests for job_sort_key and node_sort_key for a multi sched f0df380 hook fail_action 'offline_vnodes' is not functional 72405e5 Job IDs are not unique after a server restart 6526e2e Fix TPP library crash due to dangling retry packet pointer 8f21623 Mem leak fix during ralter standing resv ac3fd94 updated test_preemption of pbs_equiv_classes.py b395598 Fixed declare not found error in ubuntu and added requires bash in spec file eeca1c3 PP-810: Update cgroup hook to work with systemd ebb8382 Fix prologue hook execute twice on pbsdsh request 12fb2b0 PP-1268: Extend new windows installer to automate server/execution/client/comm config modes 81e7169 Do not power-off/ramp-down cray login nodes 494764a inbuilt hooks missing in windows 210a741 Change in PTL test to wait longer for alps cancellation 344f15a Revert leaks related to interpreter restart in suppression file 446a079 PP-832: PBSPro failover secondary server fails to continuously check whether it needs to start a scheduler locally 54fa95d Additional PTL test for PP-706

PBS Pro v18.1.3 Release

Bugs fixed in this release are as follows: 3ffce9b default_qsub_arguments -V not working (18.1) 72e45a2 Scheduler may oversubscribe node while confirming reservation

PBS Pro v18.1.2 Release Notes

Bugs fixed in this patch are as follows:

86bf85e PP-756: Upgrading from an older version to the latest mainline version fails to start the PBS daemons 51c1875 remove extra destroying of a lock in tpp