Lustre D filesystem unavailable

December 15 – 16, 2013
Conte

Update - 2:25pm, 12/16/2013

The LustreD scratch filesystem has been returned to service and both the filesystem and scheduler appear to be working properly. Conte has been returned to normal production service as of 2:20pm.

Update - 10:30am, 12/16/2013

Scheduling on Conte remains paused while engineers try to ensure the stability of the LustreD scratch filesystem. Any jobs already running should be continuing to do so.

We will send an update at 2:00pm today with further information as we work on this.

Update - 9:30pm, 12/15/2013

Storage engineers have escalated hardware issues to the storage vendor, which is analyzing log data. LustreD is currently mounted and currently-running jobs will continue, but scheduling remains paused while the issue continues to be addressed.

We will send another update by 10:00am, December 16, 2013.

Original Notice:

The Lustre D filesystem, serving the Conte cluster, has become unavailable as of about 4:30 pm Sunday 15 Dec 2013.

System engineers are working to bring the system back to 100% operation. Currently running jobs should be able to continue, but scheduling of new jobs has been disabled while the problem is being resolved.

We will update this space when we are confident the system is operating normally again.

Originally posted: December 15, 2013

Purdue University, 610 Purdue Mall, West Lafayette, IN 47907, (765) 494-4600

© 2017 Purdue University | An equal access/equal opportunity university | Copyright Complaints | Maintained by ITaP Research Computing

Trouble with this page? Disability-related accessibility issue? Please contact us at online@purdue.edu so we can help.