How To: Guides for users and Maryland T3 admins.

Help: Links and emails for further info.

Configuration: technical layout of the cluster, primarily for admins.

Log: Has been moved to a Google page, accessible only to admins.

October 2009 Log

Updated software. Discovered that OMSA regularly dies on nodes, so need to kick it regularly. Hopefully OMSA for RHEL5 won't have this problem.

October 29, 2009

MK -- Kicked OMSA on all nodes

While OMSA was still running on all nodes, it appeared to no longer be picking up system info. omreport system summary gave limited output and the HN OMSA webpage didn't show the status of anything. OMSA service starts don't appear to work completely perfectly, because I have to use an & to background the service start task - without the &, the command doesn't always return the prompt. For now, I added a notice to kick OMSA once a week to my calendar, but I need to find a method to start the service such that it runs for long periods without failure.

MK -- Changed PhEDEx configuration

srm-copy -mkdir doesn't support multiple mkdir operations. This means the .xml file passed with the -f tag must contain only one <file>...</file>. This is done in PhEDEx by setting the download-srm agent configuration -batch-files to 1 and then increasing the number of concurrent transfers using -jobs instead. Modified -batch-files to 1 and -jobs to 2.

MK -- Updated PhEDEx 3.2.9

Resumed LoadTest from FNAL, files transferred fine. Updated CVS with minor changes. Need to suspend LoadTest tomorrow after checking for failures.

MK -- Updated CRAB_2_6_3

Due to some problems with SE stageout, CRAB released new CRAB_2_6_3, so removed old directory, downloaded new CRAB_2_6_3.tgz file, and updated on interactive nodes.

MK -- Installed JDK 1.5 on some nodes

Apparently "yum update" installed new JDK releases on the interactive and grid nodes. The $JAVA_HOME environment variable was pointing to the old JDK. Force installed JDK 1.5 on the interactive and grid nodes.
Also directed users to put 'global' environment sourcing at the beginning of their ~/.chsrc & ~/.bashrc files. The OSG CE also comes with JDK releases and does set $JAVA_HOME correctly, but this environment setting was getting overwritten by sourcing the global environment last.