Maintenance was completed successfully at 2:30pm.  All previously running instances were started.

All storage nodes and clients were upgraded to gluster 3.10.0.1 and various optimizations were put into place.  Testing was performed, and a workaround was put into place for a Gluster bug.  gluster on the controller was reinstalled to fix a client corruption issue that led to high load.

 

 

On Thursday, March 23rd from 10:00am to 3:00pm all Nebula access will be shut down and unavailable.  
They will be restarted at 3:00pm. 

We will work on:

  • Software and security updates
  • Optimizations and repairs to the underlying filesystem to address stability issues
  • Stress-testing the filesystem and network

This email was sent to nebula-announce on March 16th at 11:45am:

   Nebula will take downtime for file system maintenance next week on           
   Thursday, March 23rd from 10:00 AM until 3:00 PM. During the                 
   maintenance period all instances will be paused and the Horizon              
   interface will not be available.                                             

   The downtime is necessary to resolve ongoing issues with the Gluster         
   file system on the storage nodes.                                            

   Please report any issues or concerns by either sending e-mail to             
   nebula@ncsa.illinois.edu (which will open a ticket) or writing a             
   comment in the Nebula channel of the NCSA chat.

 

 

  • No labels