[cobalt-users] RaQ550 - Mirror Drive Resync- Reboot loop

["Jeff Edwards" <jeff@xxxxxxxxxxxx>]

RE: Raq550 Resync/Reboot Issue
Hardware: Raq550, 1.0Ghz CPU, Dual 40gb, 1gb RAM (Dual 512mb Cobalt
Original), Mirror Drives
Patches: Up to date.

I have searched the archives on this issue and found numerous entries but I
though this issue was resolved by SUN long ago, But I have a server that has
lost it's mind and quickly causing me to do so also.

At 2AM the other morning, this server went into a Resync/Reboot loop as
described in previous threads on this list. The reboots originally occurred
at approx 45-50% Drive Sync.  I've tried a lot of tricks to troubleshoot
this and here is what I've found:

If I su to root and do a "kill -9 12" at some point during the early Resync,
the status of the resync will jump to ~88% and continue to completion. After
completion, the server will operate normally until either I manually reboot
or until the 4am routine daily maintenance, at which point it returns to the
Resync/Reboot loop until I run the kill command again, which buys me time
until the next 4AM window.

For some odd reason, this also caused my primary email box to become corrupt
both of the last two days shortly after 5am (maybe because my outlook "Which
runs 24x7" is checking mail during a reboot??)

I've read about skipping resync of certain partitions until others complete
and then going back, but I have idea how to manually manipulate the
Raid1syncd utilities.

I'm hesitant to Roll back the Kernel as described in messages from the June
2003 threads as SUN as released a couple of additional Kernel updated since
that period of time.

Any help, Suggestions, links, or tricks that will help me to resolve this
issue will be GREATLY appreciated!


Respectfully,


Jeff Edwards
http://www.Raq4less.com