Project updates


Jobs failed due to disk space

posted Mar 31, 2014, 11:07 AM by James Johnson

Galaxy jobs were failing in the last couple days due to a disk space issues on the queuing server.   The error message was "Cluster could not complete job".   The problem has now been resolved.

Velvet Optimizer

posted Mar 7, 2014, 9:49 AM by James Johnson

Velvet Optimizer jobs were be terminated by the queuing system for exceeding the default limit on processor use.   
New queuing params have been configured and should take effect by tomorrow.  

Monday Night Job Failures

posted Feb 18, 2014, 9:42 AM by James Johnson   [ updated Feb 18, 2014, 9:43 AM ]

A system message caused many jobs on Monday evening to be put in ERROR state.   
The problem has been corrected, and the jobs should run normally now.


Galaxy updated to release_2014.02.10

posted Feb 17, 2014, 7:02 AM by James Johnson

You should now see a "Load Data" icon to the right of the "Tools" title on the top of that panel.
This opens a file browser on your browser's host and you can specify multiple files to upload at once. 

The scratchbook is available, icon is a 3x3 matrix of tiny rectanlges just to the left of the quota meter.
With the scratchbook enabled, clicking an eye icon on a dataset will show the dataset contents in the scratch pane rather than in the center frame. 

Highlights:
  • Visualization upgrades, including Trackster CSS styling
  • Multiple Tools migrated to the Tool Shed for a leaner distribution
  • Redesign of UI rendering: new icons, new font, history pane updates
  • API functionality upgrades featuring a new master admin API key and
  • Tool Shed updates a focus on repository metadata, displays, installs, and tests

Temp Space - DRM issues

posted Feb 6, 2014, 3:17 PM by James Johnson

A number of Galaxy jobs are being terminated because they are running out of space in the temp directory.   We are attempting to set the temp dir location to alleviate that problem.   

Galaxy Maintenance Completed

posted Feb 5, 2014, 8:35 AM by James Johnson

Galaxy was moved to a new server and now uses a new job queue.   This makes more compute nodes available to galaxy.   We also adjusted application walltime and memory queuing parameters to more efficiently use our compute resources.  

Any jobs that were running at 6:00 this morning will not be recoverable, you will need to rerun them.  

Galaxy Maintenance Interruption

posted Feb 3, 2014, 1:00 PM by James Johnson

Galaxy is moving to new hardware and a new queuing server on Wednesday.  The galaxy server will go offline Tuesday evening.  Any jobs not completed by 4:00 AM Wednesday will not be recovered and will have to to be rerun.  


MSI Storage Quotas

posted Dec 2, 2013, 6:56 AM by James Johnson

The storage quota system was activated on galaxy.   You will likely see a usage meter on the right hand side of the galaxy menu bar.  The text "Using ..%" indicates the percentage of disk space you are currently using out of the maximum available to you.  Your maximum available storage is adjusted every 30 minutes to deduct space used by others working on your project within Galaxy, and space used by your project's directory in the MSI file system. If the amount of space you are using exceeds your maximum available storage, jobs you submit will not be run. 
You and other users working on your project can reclaim space by deleting and purging unneeded files or datasets.  ("Deleting" is equivalent to moving the item to "Trash/Recycle Bin", purge immediately removes the data.)  You may also want to archive histories that you are no longer actively using.  For information on archiving histories see: https://www.msi.umn.edu/sites/default/files/Galaxy_deleting_histories.pdf  
You can also email help@msi.umn.edu to request more storage. See https://www.msi.umn.edu/resources/storage for more information about MSI's storage allocation policies.

Galaxy back on line

posted Nov 26, 2013, 12:31 PM by James Johnson

The issue with failing jobs from galaxy has been resolved.   
There was an issue with communicating the exit status when the jobs were completing on the queue.


Some Galaxy jobs failing

posted Nov 26, 2013, 7:44 AM by James Johnson

A number of galaxy jobs have been failing lately, with a message in our reports of  "The cluster DRM system terminated this job".   We are trying to remedy this by allocating more memory in the job request.   

1-10 of 67