SUMMARY: Problems with Sunfire 880 picld

I originally reported:
> I'm having problems with a new install of Solaris 8 10/01 on a Sunfire
> 880.
> On reboot, we get a rash of errors, of the form:
>
> ... picld[312]: Received signal: Segmentation Fault
> ... picld[312]: WARNING: Starting up in DEGRADED mode
> 	:
> ... picld[317]: Received signal: Segmentation Fault
> ... picld[317]: Starting up in FAILSAFE mode
> ... picld[317]: Aborting - cannot recover from errors
>
> There are also subsequent errors about environmental monitoring, which
I
> haven't captured. I have only found one post that seemed related to
> picl, where the person talked about prtdiag errors in the
environmental
> monitoring section - when I run prtdiag, it just reports env.
monitoring
> as not working and goes on - presumeably due to picld failure.

I rec'd a number of responses, thanks to all who responded!
There were a variety of suggestions:
* Try a boot -r if you recently added new hardware
	[- we hadn't]
* When you install a 880 you will need to install patches:
> picl Sun-Fire-880 patch : Recommended version : 110849-10
> picl patch : Recommended version : 110460-20
	[- For us the 110849-14 patch seems to be the one that fixes
things -
	Sun insisted it, plus cluster patch kit, all be installed
	single user, and that we use absolute latest cluster patch]
* You need to have the following patches (on KR18) installed in this
order:
110460-30,110849-14,110845-03,110842-11
This is based on my recent work on 900MHz systems running 02/02 Solaris
108528-18.
	[- Sun advised 110460, 110842, and 110849, didn't mention
110845,
	but certainly these were key. The contents of 110460 and 110842
	appear to be in the most recent cluster patch, but 110849 is
not.]
* If this is v880 with 900Mhz you should try newer version
of Solaris (at least 08/02) or probably try to patch your install with
all recommended + hardware patches for v880.
	[- appears to be excellent advice, based on above. We only had
10/01,
	though, so installed that]
* I had a similar problem and wound up applying the patches below.  Hope
they help.  One is a kernel patch and is enormous.
108528-20 N.B. 21 is now out as well which I had to put on another V880
for a related problem. 110460-28 110918-04 110849-12
* One person reported seeing problems only on shutdown, not startup, so
a
non-issue for them.
	[- we saw on startup, then picld died]
* Our V880s had some type of freaky "false-fault" as a result of picld,
and in turned on all I/O fans, CPU fans, etc., thinking that there was a
problem... There is a known issue with picld and the V880s from what I
am told by engineers.  I think this is a "band-aid" until it is
ultimately solved. (Reboot required for kernel fixes, of course)... For
other fun, watch the size of picld grow to over 1 GB of memory
consumption over time.  I have to stop and restart my daemons about
every 3 weeks because they grow out of control.
	[- I haven't had them up long enough to watch for this, but
certainly
	will!]

Again, thanks!
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers