High level overview of recent events on my Server:
1. Parity Disked swapped out for larger size. Completed successfully no errors in logs.
2. Data disk (md2) swapped out for larger disk. During the rebuild, syslog noted some failed command: READ FPDMA QUEUED errors on two other data drives. Rebuild paused and server powered off to reset cables.
3. Server powered on and md1 and md6 both would not mount. However, the array autostarted and rebuild said it continued on md2.
4. Powered off server , sata cables replaced for md1 and md6 and powered on. ATA errors no longer seen in log for md1 and md6 but both will not mount. Log indicated corruptions.
5. Array put into maint. mode and xfs_repair run on both md1 and md6 successfully.
6. Array started and data rebuild for md2 continued. During this time the log begins to report XFS (md2): Corruption detected. Unmount and run xfs_repair. I let the rebuild complete.
7. After rebuild complete, ran xfs_repair -n on md2 and see the message Inode allocation btrees are too corrupted, skipping phases 6 and 7 which I have never seen before.
Can I safely run xfs_repair on md2?
Thanks for your time.
tower-diagnostics-20220417-2145.zip