Discuite: metaoffline not detach during resync

From: wwmooney@netscape.net
Date: Wed Mar 24 2004 - 13:47:18 EST


I am setting up DiskSuite 4.2.1 on a pair of SunFire 3800 systems
running Solaris 9 which are connected to a pair of A5200's.
The A5200's are fully populated with 36 GB FCAL drives, in a
split-loop configuration. luxadm shows the A5200 drives are seen
in the same order from both SF3800's and the A5200 drives have matched
major and minor numbers on both SF3800's.

The objective is to have a DiskSuite diskset of mirrored drives into
four logical volumes which would be accessed by one or the other SF 3800,
depending which system has 'taken' the diskset. This is _not_ a Cluster,
just a straight-forward sharable diskset.

Disksuite is configured on each SF 3800 with a local diskset which has
mirrored / and swap partitions. A diskset was configured on the
primary machine with eight striped metadevices: four pairs of striped
metadevices with 10, 6, 3 and 1 drives respectively. The pairs of
striped drives were configured into mirrors. A four-disk hotspare
pool was also created. The mirrors with 6, 3 and 1 drive began resyncing
immediately after the mirrors were created and the resyncing finished after
several hours.

However, the mirror comprised of the two 10-disk stripesets has been
resyncing very slowly. After 24 hours metaset shows that the mirrored
metadevice is only 18% resync'ed. The entire system (O/S, patches, etc)
was reinstalled from scratch and the same situation occurred.

The system is running SunOS 5.9 Generic_112233-11 sun4u sparc SUNW,Sun-Fire.
The patch cluster from two weeks ago has been installed. All patches
indentified by PatchPro Interactive for the hardware and software configuration have been applied. Patchdiag has been run and all patches
which related to storage had been applied. (Gotta love PatchPro and PatchDiag!)

Now I am trying to deconstruct the 10-disk stripeset to recreate it as
two smaller stripesets (on the hypothesis that the size is the problem).
But now I cannot stop the resync.

Neither 'metadetach' nor 'metaoffline' will allow me to detach the mirror
from the metadevice, not even with the -f (force) option. They fail with
    "metaoffline: hostname: array/d40 resync in progress"
Metaoffline is specificly supposed to detach when resync is in process.
Metaclear responds similiarly.

I have Google'd, RTFM'ed, etc and have not found an answer. I've used
disksuite for years and not seen this before.

A) Has anyone found a workaround for forcing the detach and deconstruction
   of the problem metadevice while it is resync'ing.

B) Has anyone seen a problem with creating >200 GB mirrors of stripesets?

I will, of course, summarize and report responses that work.

Bill Mooney
wwmooney@netscape.net
-----------------------------------------------------
Why bother working smarter, when you can work harder?

__________________________________________________________________
Introducing the New Netscape Internet Service.
Only $9.95 a month -- Sign up today at http://isp.netscape.com/register

Netscape. Just the Net You Need.

New! Netscape Toolbar for Internet Explorer
Search from anywhere on the Web and block those annoying pop-ups.
Download now at http://channels.netscape.com/ns/search/install.jsp
_______________________________________________
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 - 23:28:20 EDT