Staff have been working with assistance from NetApp support to restore service for cluster.eecs.
We are currently working on various alternatives to make the filesystems accessible again. More updates will be provided on Monday.
UPDATE
[2015-01-05 22:19:38 | Lars Rohrbach]
NetApp technical support worked with us all day, examining the problems we’re experiencing with cluster.eecs and inst-fs.eecs. We are continuing to diagnose networking issues with the fiber connection to the filer, and with router configuration.
Project shares on cluster.eecs and inst-fs.eecs remain accessible to subnets in Sutardja Dai Hall, and generally inaccessible elsewhere.
UPDATE
[2015-01-07 00:01:35 | Lars Rohrbach]
Fileservice from cluster.eecs is once again accessible.
We have made cluster.eecs available on a different subnet, using a new IP address, and both NFS and CIFS fileservice are working properly. We will continue to monitor the situation, as we work to eliminate the original routing problems.
UPDATE
[2015-01-07 12:13:02 | Lars Rohrbach]
At this time, fileservice from cluster.eecs and inst-fs.eecs is working well. Some clients may still be caching the old IP addresses, or may have stale NFS file handles — a reboot may help.
While we were experiencing problems, some project spaces were moved from cluster.eecs to project.eecs:
* demmel-d
* millennium
* netsys
* parlab
* radlab
If you utilize the above project spaces, be advised that they are now being served from project.eecs. Department NIS automount maps have been updated; if you use local mounts, you may need to change them from “cluster” to “project”.
If you are still experiencing any problems accessing your files, please try a reboot first, and then contact help@eecs if the problems continue.
Resolved as of 2015-01-06 17:00:00