Changes between Version 109 and Version 110 of PS1_IPP_Czarlog_20121203
- Timestamp:
- Dec 7, 2012, 12:28:42 PM (14 years ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
PS1_IPP_Czarlog_20121203
v109 v110 132 132 === Friday : 2012.12.07 === 133 133 * 07:20 MEH: mounts on ipp056 gone inaccessible, stalled registration. 134 * restarting rpc -- not helping134 * restarting rpc.statd -- not helping 135 135 * removing from processing 136 136 * stopping, restarting registration, stdscience, summitcopy. killing hung jobs from stack (LAPs) failed so restarted as well. … … 138 138 * 08:05 MEH: now wave4 machines are being overloaded (ipp058,054,066..). taking 1-2 instances out of processing as needed 139 139 * 08:46 EAM : rsyncs to stsci machines were causing too much load. I've stopped them, and I'll restart with lower bandwidth. 140 * 09:00 MEH: rebooting ipp056, [wiki:ipp056_log], neb-host up 140 * 09:00 MEH: rebooting ipp056, [wiki:ipp056_log], neb-host up after a bit 141 141 * 09:02 Serge: Fixed replication (see [wiki:Replication_Issues]) 142 142 * 09:05 MEH: ipp054 now the glock tarpit, has frozen registration. has lost many mounts. 143 * restarting rpc and after a bit,mounts back and registration moving along144 * then stalled again. neb-host repair143 * restarting rpc and after ~5-10 min mounts back and registration moving along 144 * then stalled again. put to neb-host repair 145 145 * 09:28 EAM : my rsync system was allowing multiple jobs to hit the same target output machine. I reworked the scripts to a single connection per source per target, so there should not be the same overloading. also re-started with 5MB/sec instead of 8MB/sec per connection 146 146 * 09:35 MEH: neb-host stsci up again, maybe help ease the load on wave4 147 * 09:45 MEH: ipp054 lost mounts to ipp030,023 again and restarting rpc not helping. rebooting [wiki:ipp054_log]147 * 09:45 MEH: ipp054 lost mounts to ipp030,023 again and restarting rpc.statd not helping. rebooting [wiki:ipp054_log] 148 148 * glocks now flushing, registration moving again 149 149 * 10:40 registration finished, slowly putting the wave4 nodes back up
