2

I have been using btrfs for over a year, and noticed reports of lots of errors in the logs but no problems. A recent scrub found HUGE numbers of csum errors, by huge I mean 85 MILLION, however scrub fixed them all. BTRFS stats reports lots of write and read io errors on 2 of the drives (zero on the third one, we are in a 3-drive raid 1.)

This would seem like a call for immediate replacement of the two drives with the errors, but there is no sign of any errors in any other logs, and smartctl reports zero raw read errors or reallocated sectors on these drives, like they are clean and happy.

It's just too many errors in btrfs to have zero from the other tools (and zero problems.) What gives? Sure, better safe than sorry, but this is 7tb to replace, possibly for no reason.

[/dev/sde2].write_io_errs   556986
[/dev/sde2].read_io_errs    4
[/dev/sde2].flush_io_errs   278
[/dev/sde2].corruption_errs 85489893
[/dev/sde2].generation_errs 0
[/dev/sdd].write_io_errs   1814862
[/dev/sdd].read_io_errs    1280445
[/dev/sdd].flush_io_errs   5327
[/dev/sdd].corruption_errs 63
[/dev/sdd].generation_errs 119
[/dev/sda2].write_io_errs   0
[/dev/sda2].read_io_errs    0
[/dev/sda2].flush_io_errs   0
[/dev/sda2].corruption_errs 0
[/dev/sda2].generation_errs 0

From sde:

  1 Raw_Read_Error_Rate     PO-R--   100   100   016    -    0
  7 Seek_Error_Rate         PO-R--   100   100   067    -    0
198 Offline_Uncorrectable   ---R--   100   100   000    -    0
199 UDMA_CRC_Error_Count    -O-R--   200   200   000    -    0

From sdd

  1 Raw_Read_Error_Rate     POSR-K   200   200   051    -    0
  7 Seek_Error_Rate         -OSR-K   100   253   000    -    0
198 Offline_Uncorrectable   ----CK   200   200   000    -    0
199 UDMA_CRC_Error_Count    -O--CK   200   200   000    -    0
200 Multi_Zone_Error_Rate   ---R--   200   200   000    -    0
  • Way old, but this kind of thing can happen with RAM errors. The scrub loads it into bad RAM, gets a checksum error and tries to fix it from mirrors. HOPEFULLY the new repaired block was calculated from a different memory area, but if not the "correction" just corrupted a bunch of data. Another reason ECC RAM is important for filesystems like BTRFS and ZFS. – Zan Lynx Jan 11 '17 at 21:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.