Cluster exceptions: Difference between revisions

Revision as of 17:54, 29 June 2026

These are exceptions to the "each cluster is separate" rule.

UID/GID is shared from Cluster 2 to Cluster 3 but not strictly enforced.
disks from Cluster 2 are exported read-only to Cluster 7 during the transition from 2 to 7.
we will experiment with SAMBA (yes, SAMBA) exports from cluster 2/7 to Cluster 8 aka coreHPC. Research only.
we tried nfs export from Cluster 2 to Cluster 3 shared/wynton. This was a disaster. Jobs running on wynton overloaded our nfs servers and made them unusable.
currently all www gateways are hosted on cluster 2. We will establish a www server on cluster 7 as soon as we can. We are also investigating hosting some of our www traffic on cluster 8.

No other exceptions known.

@@ Line 3: / Line 3: @@
 * UID/GID is shared from [[Cluster 2]] to [[Cluster 3]] but not strictly enforced.
 * disks from [[Cluster 2]] are exported read-only to [[Cluster 7]] during the transition from 2 to 7.
-* we will experiment with nfs exports from cluster 2/7 to [[Cluster 8]] aka coreHPC.  Research only.
+* we will experiment with SAMBA (yes, SAMBA) exports from cluster 2/7 to [[Cluster 8]] aka coreHPC.  Research only.
-* we tried nfs export from [[Cluster 2]] to [[Cluster 3]] shared/wynton. This was a disaster. Jobs running on wynton overloaded our nfs servers and made them unusable. This will probably happen when we export to cluster 8, but it is worth a try.
+* we tried nfs export from [[Cluster 2]] to [[Cluster 3]] shared/wynton. This was a disaster. Jobs running on wynton overloaded our nfs servers and made them unusable.
 * currently all www gateways are hosted on cluster 2. We will establish a www server on cluster 7 as soon as we can.  We are also investigating hosting some of our www traffic on cluster 8.