RE: broken disk becomes spare after reconstruct - toasters

31 May 2006


      Maintenance Garage is a new functionality (introduced in 7.0.4 I
believe) that helps to remove drives when they aren't functioning
properly and trying to recover them without outright failing them.  It
is also supposed to allow operations to continue to all other drives in
the raidgroup and when\if the drive is recoverable only update the
changes that the one questionable drive has missed.
I'm paraphrasing quite a bit, but I recall that to be the gist.  Either
way, pretty cool stuff.
Glenn
-----Original Message-----
From: Willeke, Jochen [mailto:Jochen.Willeke@wincor-nixdorf.com] 
Sent: Wednesday, May 31, 2006 9:51 AM
To: Glenn Walker; toasters@mathworks.com
Subject: RE: broken disk becomes spare after reconstruct
Hi toasters,
we are running 7.0.1R1.
BTW: what is the maintenance garage? Never heard of this.
Regards
Jochen
-----Original Message-----
From: Glenn Walker [mailto:ggwalker@mindspring.com] 
Sent: Wednesday, May 31, 2006 3:44 PM
To: Willeke, Jochen; toasters@mathworks.com
Subject: RE: broken disk becomes spare after reconstruct
Basically it appears that the disk started to work correctly again (ie,
it was responsive to commands).  At that point, it will return it to
spare pool...
Most of the time disks don't actually physically fail - ONTAP is very
concerned about data consistency so it will fail misbehaving drives so
that it no longer uses them.  Sometimes the drives start behaving again,
and can be used...
ATA drives are very bad for this:  they typically go into internal
recovery procedures and don't communicate status to the controller\OS -
FCAL drives are very good about communicating status and continuing with
other operations in the meantime.  I've typically seen this behavior on
R200's more than FAS-series (with FCAL drives).
Out of curiosity - are you utilizing 7.0.4 or newer?  Wondering if
maintenance garage would help...
Glenn
-----Original Message-----
From: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com]
On Behalf Of Willeke, Jochen
Sent: Wednesday, May 31, 2006 8:50 AM
To: toasters@mathworks.com
Subject: broken disk becomes spare after reconstruct
Hello Toasters :-)
i have seen something strange on one of our R200. Last week i got an
error message with the following output:
[scsi.cmd.notReadyCondition:notice]: Device 2b.35: Device returns not
yet ready: CDB 0x28:18974bb9:0009: Sense Data SCSI:not ready -  (0x2 -
0x4 0x7 0x2a)(65103).
[scsi.cmd.checkCondition:error]: Device 2b.35: Check Condition: CDB
0x28:154efb9d:022e: Sense Data SCSI:not ready -  (0x2 - 0x4 0x7
0x2a)(209311).
[raid.config.filesystem.disk.recovering:error]: Attempting to bring file
system Disk /aggr0/plex0/rg1/2b.35 Shelf 2 Bay 3 [NETAPP
X266_MTOMC320PAD R5VV] S/N [A81K4MBE] back into service.
......normal procedure of taking one spare and reconstruct the raid
group. But then comes what i am wondering about.....
[raid.assim.disk.spare:notice]: Sparing Disk /2b.35 Shelf 2 Bay 3
[NETAPP   X266_MTOMC320PAD R5VV] S/N [A81K4MBE], because volume is
online and complete
The System takes the broken disk as a spare?!? How can this be?
Perhaps anybody can help because NOW does not explain this situation
deep enough to make it understandable to me :D
Regards
Jochen