7

I'm maintaining a pre-configured storage system for a project team. The storage consists of 36 x 8 TB HGST Disks (~240 TB net) in RAID60 (2x18 disks in RAID6), Controller is a MegaRAID SAS 9361-8i, OS=Scientific Linux 7.2

storcli64 /c0 /v1 show
DG/VD TYPE   State Access Consist Cache Cac sCC       Size Name
0/1   RAID60 Optl  RW     Yes     RWBD  -   ON  232.860 TB      

When running weekly scheduled consistency checks (or if I trigger them manually), I get a lot of inconsistency warnings, but they never seem to be fixed - during the next cc the same warnings reappear:

[root@faith ~]# grep 22649241 /var/log/message*
/var/log/messages:Oct 25 20:54:01 faith MR_MONITOR[1604]: <MRMON063> Controller ID:  0   Consistency Check found inconsistent parity on VD#012    strip:  #012    ( VD   =   1,   strip#012      =   22649241)
/var/log/messages-20161002:Sep 25 22:34:03 faith MR_MONITOR[1604]: <MRMON063> Controller ID:  0   Consistency Check found inconsistent parity on VD#012    strip:  #012    ( VD   =   1,   strip#012      =   22649241)
/var/log/messages-20161024:Oct 23 22:35:46 faith MR_MONITOR[1604]: <MRMON063> Controller ID:  0   Consistency Check found inconsistent parity on VD#012    strip:  #012    ( VD   =   1,   strip#012      =   22649241)

What would be the best way to get array consistent? I guess a rebuild could likely fail in the current state? Or might this be a side effect of the raid60?

  • 3
    What's "a lot"? Ten? Ten thousand? Do the same strip values ever repeat? And putting 18 8TB disks in RAID6? I hope that data isn't important, or you have good backups - and a lot of time to restore. – Andrew Henle Oct 25 '16 at 21:59
  • The eventlog reports (always) 24 corrections, I checked a few values and they repeat constantly in every consistency check from the very first (e.g. strip 199999 -> grep 199999 event.log | wc -l = 29). Maybe the cc is confused by the raid60 (since a cc couldn't apply for the raid0 component) - but I assume the cc runs on every raid6 array, not on the raid60 disk group? – alphabricks Oct 26 '16 at 6:49
  • 1
    Is the HBA firmware up-to-date? See avagotech.com/support/download-search I'd also dig through the support docs there, looking for known issues. These kind of issues are one technical reason why high-end storage systems tend to use only vendor-supplied disks of limited variety instead of allowing you to plug in any old disk. Somebody has to spend the time to make sure the entire system works for those customers willing to pay for "It WILL work" for systems storing their critical data. – Andrew Henle Oct 27 '16 at 10:15
  • @ Andrew Henle - no, it's just marketing dealing. But using disks not designed for RAID in a RAID can have drawbacks. – Overmind Jan 4 '18 at 7:38

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.