Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Lars Ellenberg wrote: >>>I think that neither of these can be confirmed. >>>If we mount the filesystem on /dev/md3, and export over NFS, everything >>>works like a charm. No problems whatsoever. All with full NFS load. >>>(But without traffic over Gigabit crossover link, I must admit). Now that think about it, drbd traffic was actually *not* involved when the error occurred last time. Drbd was in WFConnection mode from the beginning, the other machine did not run drbd. So, the difference is just "fs on standalone drbd on md" versus "fs on md" - first gives error and the second works. I would suspect that drbd touches some execution path in the underlying driver that would not be run by filesystem code. And triggers a bug there. Does it make sense? Note however that there is an 'md' layer on top of hardware driver in both "good" and "bad" cases... Eugene