> From: "Roy G. Culley" <[EMAIL PROTECTED]>
>
> On two occasions we have had our Sunscreen EFS3.0 cluster
> systems (we have 2 such clusters and this problem has
> shown up once on each cluster) hang with the following
> messages being sent to the console:
>
> qfe0 allocb fail lock_set_spl lock held and only one cpu
>
[...]
> break to the serial port. Does anyone know why this happens as
> it defeats the whole HA concept if both nodes in the cluster
> fail together?
>
I have not seen this problem before specificly on SunScreen. I would
be interested in knowing which kernel and qfe patches you have applied
to the system. If not related to the kernel or quad cards, then we
should investigate this further.
> As an aside is there any command to get info / statistsics
> regarding state tables used by the Sunscreen EFS3.x? For the
> spf200 I could:
>
> ss_client sunscreen stats
>From a 3.x admin station, you can run (after logging in via CLI):
ssadm -r <screenname> lib/support stats
To see a listing of other "support" commands, you can run:
ssadm -r <screenname> lib/support help
There is a "Command Compatibility Reference Table" on page 171
of the Reference Manual that will tell you what the equivalent
3.x commands are for old SPF-200 and EFS 1.1/2.0 commands (but
"stats" is actually missing from that list - oops!).
>
> By grep'ing for max_entries I could see how many slots were being
> used.
FYI: SunScreen EFS 3.0 no longer enforced the max_entries variable, and
the variable was completely removed in SunScreen 3.1.
hope this helps,
Valerie
SunScreen Developer
[EMAIL PROTECTED] -or- [EMAIL PROTECTED]
-
[To unsubscribe, send mail to [EMAIL PROTECTED] with
"unsubscribe firewalls" in the body of the message.]
