IBM Platform LSF 10.1.0.8 Fix 529611 Readme

Abstract

This package includes new support for MS Azure Cycle Cloud in LSF Resource Connector which is included in LSF Fix Pack 9 (10.1.0.9). This fix also includes Fix 528785 to address issues when cloud hosts joining the LSF cluster reuses the same host name.

Description

Readme documentation for IBM Platform LSF Fix 529611 including installation-related instructions, prerequisites and co-requisites, and list of fixes.

This fix addresses the following issues:
- New LSF Resource Connector provider to work with Microsoft Azure Cycle Cloud
- Multiple fixes for 528785 to address issues when cloud hosts joining LSF cluster reuses the same host name. When cloud bursting is enabled with LSF resource connector, cloud hosts joining LSF cluster might reuse the same host name. This fix addresses multiple issues with the frequent reuse of host names when different hosts join then LSF cluster, which might cause jobs to dispatch to the wrong host. In this fix, LSF ensures that local resource strings are updated correctly when host names are reused frequently, and jobs are dispatched to the correct hosts as expected.

Readme file for: IBM® Spectrum LSF
Product/Component Release: 10.1.0.8
Update name: Fix 529611
Fix ID: LSF-10.1-build529611
Publication date:19 September 2019
Last modified:19 September 2019

Contents:

• List of fixes
• Download location
• Products or components affected
• System requirements
• Installation and configuration
• Uninstallation
• List of files
• Copyright and trademark information

• List of fixes

APAR#: P103201

• Download Location

Download Fix 529611 from the following location: http://www.ibm.com/eserver/support/fixes/

• Products or components affected

Affected components include:

- Master lim

- Slave lim

- mbatchd host management handling

- Ebrokerd

• System requirements

x86-64 Platform

• Installation

Refer to IBM Knowledge Center for the instation steps of LSF Resource Connector for Cycle Cloud

Note: This patch includes changes and new files to $LSF_ENVDIR. Make sure to manually copy the conf/ files to the directory as patchinstall does not automatically update $LSF_ENVDIR.




The included fix, 528785, requires LSF 10.1.0.8 with resource connector functionality and updated user_data.sh startup script to include instanceID resource in the slave's local resource.

5.1 Before installation

• 1. Close all hosts and inactivate all queues
  

  badmin hclose all
  badmin qinact all

• 2. Wait for approximately 10 minutes to let all previous machine requests finish
• 3. Update slave host startup script with the script that includes instanceID resource in the LSF_LOCAL_RESOURCE entry in the slave's lsf.conf file


• 4. Configure lsf.shared file to include "instanceID" string for LIM to accept instanceID resource from slave hosts if not already configured
  

  In lsf.shared under Resource section
  instanceID String () () (unique id of machines)

5.2 Installation steps

• 1. Install the new patch from the install directory under ${LSF_TOP}/10.1/install
  

  ./patchinstall <patch>

• 2. Make sure file permissions are correct for the file ${LSF_TOP}/work/<clustername>/resource_connector/hosts.json file are set to the correct user
  

  -rw-r--r-- 1 lsfadmin lsf 2135 Sep 4 12:26 hosts.json

• 3. Ensure all slave hosts are updated also. If using VM images, make sure all images are updated and saved.

5.3 After installation

• 1. Shutdown and startup all lsf daemons
  

  badmin hshutdown all
  lsadmin resshutdown all
  lsadmin limshutdown all
  

  lsadmin limstartup all
  lsadmin resstartup all
  badmin hstartup all

• 2. Open all hosts and activate all queues
  

  badmin hopen all
  badmin qact all

• Uninstallation

• 1. Uninstall from the install directory under ${LSF_TOP}/10.1/install, make sure you do the same for the slave hosts
  

  ./patchinstall -r <patch>

• List of files

This Fix includes the following files.

• - ebrokerd
• - lim
• - mbatchd
• - mbschd
• - schmod_demand.so
• - example_hostProviders.json
• - cyclecloud/conf (new directory)
• - lsf-headless-cyclecloud.txt
• - cyclecloudprov_config.json
• - cyclecloudprov_templates.json
• - cyclecloud/lib (new directory)
• - commons-collections-3.2.2.jar
• - commons-io-2.5.jar
• - commons-lang-2.6.jar
• - CycleCloudTool.jar
• - jackson-annotations-2.9.8.jar
• - jackson-core-2.9.8.jar
• - jackson-databind-2.9.8.jar
• - log4j-api-2.8.2.jar
• - log4j-core-2.8.2.jar
• - okhttp-3.14.1.jar
• - okio-1.17.2.jar
• - cyclecloud/scripts (new directory)
• - getAvailableMachines.sh
• - getAvailableTemplates.sh
• - getRequestStatus.sh
• - getReturnRequests.sh
• - requestMachines.sh
• - requestReturnMachines.sh
• - user_data.sh

• Copyright and trademark information

© Copyright IBM Corporation 2019

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