Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Our problem was a sudden system load that bogged down our DRBD servers for 13-14 hours daily. The problem turned out to be in the batteries on the RAID controllers. When the batteries are low and unable to fully charge, they force the RAID system into write-through mode instead of write-back mode while they attempt to re-charge/adjust. This slowed down everything and caused our mysterious increased system load. Once we replaced the batteries, system load returned to normal. We still have one drive doing 10 times the writes as the other, but performance is no longer affected, so finding the cause of that problem is no longer as critical (but we continue to investigate). I would like to thank those on the list who tried to help, we really do appreciate the effort. Thanks again. - Richard