Readme and Release notes for release 4.1.0.7 (LL) IBM Tivoli Workload Scheduler LoadLeveler 4.1.0.7 LL_scheduler-4.1.0.7-power-AIX Readme

Unless specifically noted otherwise, this history of problems fixed for LoadLeveler 4.1.0.x applies to:

• LoadLeveler 4.1.0.x for AIX 6
• LoadLeveler 4.1.0.x for AIX 5
• LoadLeveler 4.1.0.x for SUSE LINUX Enterprise Server 11 (SLES11) on POWER servers
• LoadLeveler 4.1.0.x for Red Hat Enterprise Linux 5 (RHEL5) on servers with 64-bit Opteron or EM64T processors
• LoadLeveler 4.1.0.x for Red Hat Enterprise Linux 5 (RHEL5) on Intel based servers

For LL 4.1:

• On any machine where you plan to install the scheduler rpm, you must install the resource manager rpm and the resource manager rpm must be installed before the scheduler rpm.
• If the scheduler and resource manager filesets are installed on the same machine, then the LoadLeveler version of those filesets have to be at the same level in order for transactions to be processed correctly.

For LL 4.1.0.1:

• Machine group support is limited to file-based configuration. Fixed in LL 4.1.0.2.
• The "Reconfigure" button in the configuration editor does not reconfigure the nodes. When you have completed your configuration editor updates, run "llctl -g reconfig" on your LoadLeveler cluster to ensure all daemons read the updates. Fixed in LL 4.1.0.2.

For LL 4.1.0.5+:

Problems fixed in LoadLeveler 4.1.0.7 [August 6, 2010]

• Fixed schedd crash due to formatted message less than buffer size when running fairshare.
• Fixed SEGV abort in the schedd and startd daemons due to failures when writing to savelog directories.
• Fixed jobs in hold state due to error in cleaning up switch tables.
• Resource Manager only:
  • Fixed jobs staying in ST state when running heavy preemption load due to race condition in the reject and ready status notification when disabling windows.

Problems fixed in LoadLeveler 4.1.0.6 [June 11, 2010]

• Fixed the consumable CPUs to be evaluated correctly for jobs which dynamically turn smt on and off so jobs will be scheduled properly on Power5 or Power6 systems.
• Scheduler only:
  • A design change was implemented so that dependent steps are also counted as "queued" steps for the purposes of enforcing maxqueued and maxidle limits, and so a dependent step which is at the maxqueued limit will get a new qdate.
  • Linux: Fixed LoadLeveler to be able to change the duration of an active reservation of a Blue Gene Partition on the BG/P system.

Problems fixed in LoadLeveler 4.1.0.5 [May 5, 2010]

• Fixed multiple llconfig issues.
• Fixed region manager coredump due to machine locking issue in the base library.
• Fixed LoadLeveler to handle 0 in the IP address octets correctly plus fixed potential memory error when handling addresses.
• Scheduler only:

  Linux: Fixed LoadLeveler to skip over invalid partition in the BlueGene database during partition loading so that jobs can start on the free nodes.

Problems fixed in LoadLeveler 4.1.0.4 [March 15, 2010]

• Fixed LoadLeveler jobs to start correctly in a login shell and fixed the pmd from hanging during execution.
• Scheduler only:
  • Fixed the reservation debug message field so the central manager will not core dump.
  • Fixed the dispatch cycle of routed jobs so when the central manager failover takes place, the preempted jobs will now be able to run.

Problems fixed in LoadLeveler 4.1.0.3 [February 22, 2010]

• Fixed the llsummary command segmentation fault by skipping over data that are not valid when processing the history file.
• Fixed LoadLeveler to not send notification emails if the api process has already reported the errors.
• Fixed LoadLeveler jobs to run with the correct gid on AIX platform.
• Resource Manager only:
  • Fixed LoadLeveler jobs from hanging in preempted pending state by correcting the machine state for the jobs being preempted.
  • Fixed the schedd daemon memory leak when processing reservations by removing the reservation element object after use.
  • Fixed LoadLeveler to set the environment variable from the prolog output if each line contains at most 65534 characters. All lines containing more than 65534 characters will be ignored.
• Scheduler only:
  • Fixed LoadLeveler to execute the rescan function so jobs can not be scheduled once the running jobs are completed when using the default scheduler.
  • Fixed the negotiator daemon to ensure that resource counts are now being updated correctly when a step is canceled during the window of time after it has been scheduled but before the job start order has been dispatched.

Problems fixed in LoadLeveler 4.1.0.2 [December 21, 2009]

• Fixed user id size so that jobs submitted by user id name up to a length of 256 can run.
• Fixed the llqres command to display the correct number of days for months that are shorter than 31 days under the Recurrence output section.
• Fixed submission so that jobs will be rejected when user id is invalid.
• Fixed LoadLeveler to not core dump when started in drain mode.
• Fixed LoadLeveler multi-step jobs to run with the correct umask.
• Fixed sending of emails to the correct LoadLeveler admininstrators when route failure email error is detected.

Problems fixed in LoadLeveler 4.1.0.1 [November 23, 2009]

• Fixed LoadLeveler performance by reducing the overhead of handling llq query requests so that the impact to the overall scheduling progress is also reduced.
• Fixed core dumps, memory and deadlock issues in different LoadLeveler daemons.

Problems fixed in LoadLeveler 4.1.0.1 [November 23, 2009]