68 lines
3.3 KiB
Plaintext
68 lines
3.3 KiB
Plaintext
TORQUE Resource Manager 5.1.0 Release Notes
|
||
January 2015
|
||
|
||
The release notes file contains the following sections:
|
||
|
||
> New Features
|
||
> Differences
|
||
> Known Issues
|
||
> Resolved Issues
|
||
|
||
===== New Features =====
|
||
|
||
The following is a summary of key new features in TORQUE Resource Manager 5.1.0.
|
||
|
||
==> Ability to Provide Condensed qstat Output
|
||
|
||
- A 'qstat -C' option, which specifies that TORQUE will provide only a condensed output (job name, resources used, queue, state, and job owner) for jobs that have not changed recently (as per the job_full_report_time parameter), has been added. Jobs that have recently changed will continue to send a full output.
|
||
|
||
==> Performance Enhancements to MOM Clean-up Time
|
||
|
||
- Some minor performance enhancements were made to improve MOM clean-up time.
|
||
|
||
===== Differences =====
|
||
|
||
This section contains differences in previously existing features that require a change in configuration or routine.
|
||
|
||
==> CLIENTRETRY Configuration Option Support for qdel
|
||
|
||
- The qdel parameter now includes the -b option and CLIENTRETRY configuration option support. This feature functions similar to -b option for qsub. Specifically:
|
||
* -b <num> command line argument support
|
||
* CLIENTRETRY configuration option support
|
||
* PBS_CLIENTRETRY environment variable support
|
||
|
||
==> Jobs Deleted When the Dependency Can No Longer Be Satisfied
|
||
|
||
- When a job is deleted because its dependency can no longer be satisfied, that job will follow the keep_completed parameter set (if any) for pbs_server. Previously, jobs were purged immediately.
|
||
|
||
==> pbs_server -t No Longer Supports hot|warm|cold Options
|
||
|
||
- The pbs_server -t option no longer supports the hot|warm|cold Options. Other options are still supported.
|
||
|
||
===== Known Issues =====
|
||
|
||
The following are known issues in the TORQUE Resource Manager 5.1.0.
|
||
|
||
> DNS caches are not purged of removed nodes when Elastic Computing is enabled. SLES 11 SP1 has an issue with giving the old IP address toTORQUE even after updating /etc/hosts. Do not use SLES SP1 if you are using this method to manage the IP addresses for pbs_server. The mom'sOSs are irrelevant. (TRQ-2765)
|
||
|
||
> Some limitations exist in the way that pbsdsh can be used. Please note the following situations are not currently supported:
|
||
- Running multiple instances of pbsdsh concurrently within a single job.(TRQ-2851)
|
||
- Using the -o and -s options concurrently; although requesting these options together is permitted, only the output from the first node is displayed rather than output from every node in the chain. (TRQ2690)
|
||
|
||
|
||
===== Resolved Issues =====
|
||
|
||
The following is a list of some key bugs fixed in TORQUE Resource Manager 5.1.0. Following each issue description is an associated issue number in parentheses.
|
||
|
||
> HOST_NAME_SUFFIX was no longer adding suffix to job names. This bug is fixed. (TRQ-2956)
|
||
|
||
> pbs_mom filling up the logs in a HA environment. Reduced verbosity in error logging in HA environments. (TRQ-2863)
|
||
|
||
> Delays reported with Preemption. The following changes were made to reduce delays: (TRQ-3026)
|
||
- Speed up process data collection to help diminish time to rerun a job.
|
||
- Correctly sum job memory and cpu time resources for processes created under setsid() calls within a job.
|
||
- Avoid an infinite loop in pbs_mom when running under a LinuxContainer (lxc).
|
||
|
||
|
||
|
||
Copyright 2015, Adaptive Computing Enterprises, Inc. |