SDS Mirror Problem

From: Adam Ronthal (aronthal@cirba.com)
Date: Tue Jun 10 2003 - 11:17:25 EDT


Hi all...

Not quite sure how to resolve this one, but here are the symptoms:

We are running SDS 4.2.1 with mirrored disks in a RAID 1+0 configuration.

/var/adm/messages is showing:

Jun 9 15:14:23 osftserv02 md_stripe: [ID 641072 kern.warning] WARNING:
md: d11: read error on /dev/dsk/c4t36d0s6

.... which appears to indicate that we're losing a disk.

The system was rebooted and the /dev/md/rdsk/d1 was fsck'd (which was
probably a bad idea, as now there is no data in it).... however....

We ran a metareplace:

metareplace d1 c4t33d0s6 c4t42d0s6

which set that stripe to resync itself.... but the resyncing has been
running overnight and shows no sign on completing.

The output of metastat shows:

d1: Mirror
     Submirror 0: d10
       State: Needs maintenance
     Submirror 1: d11
       State: Needs maintenance
     Pass: 1
     Read option: roundrobin (default)
     Write option: parallel (default)
     Size: 106029424 blocks

d10: Submirror of d1
     State: Needs maintenance
     Invoke: metareplace d1 c4t34d0s6 <new device>
     Size: 106029424 blocks
     Stripe 0: (interlace: 32 blocks)
         Device Start Block Dbase State Hot Spare
         c4t32d0s6 0 No Maintenance
         c4t42d0s6 0 No Resyncing
         c4t34d0s6 0 No Maintenance

d11: Submirror of d1
     State: Needs maintenance
     Invoke: after replacing "Maintenance" components:
                 metareplace d1 c4t36d0s6 <new device>
     Size: 106029424 blocks
     Stripe 0: (interlace: 32 blocks)
         Device Start Block Dbase State Hot Spare
         c4t35d0s6 0 No Last Erred
         c4t36d0s6 0 No Last Erred
         c4t37d0s6 0 No Last Erred

d2: Mirror
     Submirror 0: d20
       State: Needs maintenance
     Submirror 1: d21
       State: Needs maintenance
     Pass: 1
     Read option: roundrobin (default)
     Write option: parallel (default)
     Size: 70684712 blocks

d20: Submirror of d2
     State: Needs maintenance
     Invoke: after replacing "Maintenance" components:
                 metareplace d2 c4t39d0s6 <new device>
     Size: 70684712 blocks
     Stripe 0: (interlace: 32 blocks)
         Device Start Block Dbase State Hot Spare
         c4t38d0s6 0 No Maintenance
         c4t39d0s6 0 No Last Erred

d21: Submirror of d2
     State: Needs maintenance
     Invoke: metareplace d2 c4t41d0s6 <new device>
     Size: 70684712 blocks
     Stripe 0: (interlace: 32 blocks)
         Device Start Block Dbase State Hot Spare
         c4t40d0s6 0 No Last Erred
         c4t41d0s6 0 No Maintenance

So the main questions are:

1) Is our d1 partition lost? (we can restore from backups)
2) How do I interpret the state of the d2 mirror and how do I recover from it?

Thanks in advance, will summarize to the list.

-Adam
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers



This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:26:33 EDT