IBM Spectrum LSF RTM 10.1.0 Fix Pack 4 (473431) Readme File
Abstract
This Fix Pack includes 20 fixes and 8 enhancements since the release of IBM Spectrum LSF RTM 10.1.0.3 on April 28, 2017.
Description
Readme documentation for IBM Spectrum LSF RTM 10.1.0 Fix Pack 4 (473431) including installation-related instructions, prerequisites and co-requisites, what's new and list of fixes.
Details for the enhancements are listed in
the section "1.What's new".
Details for the fixes are listed in the section "2. List of fixes".
Details for new Known Issues and Limitations are listed in the section
"3. Known Issues and Limitations".
Readme file for: IBM® Spectrum LSF RTM
Product/Component Release: 10.1.0
Update Name: Fix Pack 473431
Fix ID: rtm10.1.0-build473431
Publication date: 27 October 2017
Last modified date: 27 Octoberl 2017
Contents:
1. What's new
2. List of fixes
3. Known Issues and Limitations
4. Download location
5. Products or components affected
6. System requirements
7. Installation and configuration
8. List of files
9. Product notifications
10. Copyright and trademark information
1. What's new
* New Lichist (License History) Plugin added, with the following capabilities:
a, Captures historical license usage.
b, Calculates the "conflicts" that occur when multiple jobs from the same user are executing concurrently on the same first execution host. Since lmstat only returns the user and hostname information, we cannot
always determine which specific LSF job (if any) checked out a license on a given host.
c, An additional tab (Grid > Job Details > License Usage) to display both the current and historic license usage for a job, and the calculated conflicts for that job.
d, An additional table on page Grid > By host > View Host Job Detail that displays both the current and historic license usage for a host.
* Changes to GRID graph templates:
a, To accommodate the difference between Job and Job Slot usage (and to display data for both), the 'GRID - Projects - Level 1 - Running Jobs’ and ‘GRID - Projects - Level 1 - Pending Jobs’ graph templates
(found under the Console tab - Templates > Graph Templates) have been renamed to 'GRID - Projects - Level 1 - Running Job Slots’ and ‘GRID - Projects - Level 1 - Pending Job Slots’, respectively.
b, Two new graph templates have been added to replace the previous graph templates. The new 'GRID - Projects - Level 1 - Running Jobs’ graph template shows the running jobs by project for the cluster.
The new ‘GRID - Projects - Level 1 - Pending Jobs’ graph template shows the pending jobs by project for the cluster.
c, Columns for Total Jobs, Pending Jobs, and Running Jobs have also been added to the Job Info > By Project screen in the Grid tab if check on in the Settings.
* Ability to map project name to predefined string. The file
projmap/setup.php.demo has been provided to demostrate the usage.
* The License Scheduler 10.1 FEATURE_DELTA data in will be collected in
the database by x86_64 poller of LSF RTM (when FEATURE_DELTA is enabled in License Scheduler
10.1).
* Change to use a short hostname in RTM to avoid performance issues in
large installations. A script has been provided
(database_shorten_hostname.sh) to shorten the host name in the RTM
database. The script is
called by rtm_patch.sh during Fix Pack 4 installation. We strongly suggest using short hostnames (without domains). It may be run later (after Fix Pack 4 installation), but must be run before future Fix Packs (that is,
Fix Pack 5).
* Enhanced Daily Stats and Statistical Dashboard. In the Daily Statistics
(Grid > Reports > Daily Statistics) page, the ‘timespan’
filter has been changed to be more flexible and a ‘Unit’ filter has been added for the time
column (including Minutes, Hours, Days, Weeks and Months). In the Statistical Dashboard (Grid > Dashboards > Statistical), the ‘Views’ filter has been changed from multiple checkboxes to a dropdown list and User,
Queue, Project, Unit, and timespan filters have been added..
* Collection of GPU data from LSF 10.1 API. This includes an additional
4 graph templates for the ELIM portion of the GPU data:.
- GPU Utilization (Utilization of the first four GPUs on the host)
- GPU Memory Utilization (Memory Utilization of the first two GPUs and total GPUs on the host)
- Shared GPU Utilization (Utilization of the GPUs in share mode on the host)
- Shared GPU Memory Utilization (Memory Utilization of the GPUs in share mode on the host).
* Added 12 graph templates for perfmon (LSF Performance Matrix):
GRID - LSF Host Info Requests
GRID - LSF Host Match Criteria
GRID - LSF Job Buckets
GRID - LSF Job Info Requests
GRID - LSF Job Scheduling Interval
GRID - LSF Job Submit Requests
GRID - LSF Jobs Completed
GRID - LSF Jobs Dispatched
GRID - LSF Jobs Submitted
GRID - LSF MBatchD Requests
GRID - LSF MBD File Descriptor Usage
GRID - LSF Queue Info Requests
2. List of fixes
ID |
Description |
168486 |
Fix for the grid_add_graphs.php taking more than 40 minutes to add licenseProject graph. |
167993 |
Fix for the issue in the Input String of the 'GRID - Cluster Pending by Pending Reason' Data Input Method. |
167528 |
Fix for the gap issue in job graph. |
166679 |
Merge two CVE fixes, Security Vulnerabilities in cacti 0.8.7g. |
164531 |
Fix for the parsing error when blstat output has a new line for FEATURE_DELTA. |
163855 |
Fix for the rtmssh issue with JAVA 8 security. |
163632 |
Fix for the issue where two reason codes (11701 and 2601) are given for the same pending reason. |
156893 |
Fix for the incorrect ELIM value when LSF_UNIT_FOR_LIMITS=GB is set in the LSF cluster. |
156568 |
Fix for the incorrect error messages given for licflexpoller that mislead users. |
152840 |
Fix for inability to import xml file with special characters. |
149196 |
Fix for the issue where Job graphs do not match the selected Time Zone. |
171099 |
Fix for the extra records to get purged from the grid_jobs_jobhosts_finished table for reused jobids. |
163487 |
Fix for a SQL error when the description of an LSF parameter is too long (that is, more than 1024 Bytes). |
161366 |
Add validation for the path of license poller. |
158923 |
Fix for the issue where the wrong pending reason (r15s) are given for non-resource related issues with the suspend subreason. |
158862 |
Fix for long fix pack upgrade time issue on large installations. |
157356 |
Fix for the issue where the license message cannot be retrieved from License Scheduler when LSF uses an alias cluster name. |
157901 |
Removal of the 'All' option on page Job Details > Settings > Rows to Display to avoid crashing the RTM Server in larger installations. |
155168 |
Fix for issue where RTM on RHEL7 fails to deploy when the user configures an extra repo RHN "optional" channel. |
155034 |
Fix for the incorrect unit used for Temp/Swap when LSF_ENABLE_TMP_UNIT is enabled in the LSF cluster. |
3. Known Issues and Limitations
Issue |
Description |
If the host name contains a period, RTM cannot identify the domain name and gets an incorrectly shortened name with run database_shorten_hostname.sh |
For example, if the host name is abc.dd (FQDNs is abc.dd.domain.com), the shorten hostname script will get the domain name as dd.domain.com, and host name as abc, not abc.dd. |
4. Download Location
Download Fix Pack 473431 from the following location: http://www.ibm.com/eserver/support/fixes/
5. Products or components affected
Components affected by the new issues
addressed in IBM Spectrum LSF RTM Version 10.1.0 Fix Pack 4 include:
Grid Plugin
Disku Plugin
License Plugin
License Scheduler Plugin (Gridblstat)
Heuristics Plugin (JobIQ)
Meta Plugin
Benchmark Plugin
Gridalarms Plugin
Rtmssh Plugin
6. System requirements
OS of RTM master:
RHEL6.2+ x86_64, RHEL7.1+ x86_64, RHEL7.2 on Power 8 Little Endian and SLES11.1~11.4 x86_64
7. Installation and configuration
7.1 Before installation
Download the patch file. Follow below steps to install.
1) Login to RTM master host as root
2) #tar xzf rtm10.1-build473431-xxx-xxx.tar.gz
If you installed IBM Spectrum LSF RTM 10.1.0, and performed some customization, back up your customizations before you install Fix Pack 4 . Otherwise, your customizations will be lost.
Note:
a, Before RTM 10.1 GA fresh install on SLES 11.4, follow below steps:
Extract rtm_install.sh from rtm10.1.0-build473431-sles11-x64.tar.gz, then use it to replace rtm_install.sh under RTM 10.1 GA install package location.
b, If you configured the extra repo RHN "optional" channel on RHEL7, and install RTM 10.1 GA on non-default path (not /opt/IBM/). follow below steps:
Extract rtm_install.sh from rtm10.1.0-build473431-rhel6-x64.tar.gz, then use it to replace rtm_install.sh under RTM 10.1 GA install package location.
c, For installing FixPack 4 on RHEL 7 x86_64, use rtm10.1.0-build473431-rhel6-x64.tar.gz package.
7.2 Installation steps
1) Run
rtm_patch.sh as below,
#./rtm_patch.sh
Note:
a, If you need to delopy the patch on a remote poller, install the rpm package of poller. For example:
#cd x86_64
#rpm -Uvh --replacefiles --nodeps --prefix $RTM_TOP rtm-poller-10.1.0-4.473431.x86_64.rpm rtm-client-10.1.0-4.473431.x86_64.rpm
b, If the eauth patch has been applied to the LSF cluster that RTM monitors, replace the old version of eauth in RTM_TOP/rtm/lsf<8/9/10>/etc/eauth with the new eauth.cve on the RTM server. For example:
#mv /opt/IBM/rtm/lsf101/etc/eauth /opt/IBM/rtm/lsf101/etc/eauth_bak
#mv /opt/IBM/rtm/lsf101/etc/eauth.cve /opt/IBM/rtm/lsf101/etc/eauth
7.3 After installation
N/A
7.4 Uninstallation
All files changed with this Fix Pack are automatically backed up as a tar file patch_bak<time_stamp>.tar.gz under $RTM_TOP. To roll back this patch, restore all files included in the tar file.
8. List of files in package
cacti/include/config.php
9. Product notifications
To receive information about product solution and patch updates automatically, subscribe to product notifications on the My notifications page (www.ibm.com/support/mynotifications) on the IBM Support website (support.ibm.com). You can edit your subscription settings to choose the types of information you want to get notification about, for example, security bulletins, fixes, troubleshooting, and product enhancements or documentation changes.
10. Copyright and trademark information
© Copyright IBM Corporation 2017
U.S. Government Users Restricted Rights - Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM Corp.
IBM®, the IBM logo and ibm.com® are trademarks of International Business Machines Corp., registered in many jurisdictions worldwide. Other product and service names might be trademarks of IBM or other companies. A current list of IBM trademarks is available on the Web at "Copyright and trademark information" at www.ibm.com/legal/copytrade.shtml.