🚀 go-pugleaf

In article <1992Oct15.093117.1009@beckman.com>, dsroberts@beckman.com writes:
> I'm having a problem with ANALYZE (/system and /crash) on my cluster.  From the
> same username (whether my own or SYSTEM), it works fine on the 6610 and 6410,
> but from the 6420 I get "virtual memory exhausted" at different points in the
> analyze.  Sometimes it occurs right away, sometimes after a read/exec,
> sometimes in the middle of show stack.  I'm guessing that process quota's
> aren't the problem, since I don't have this problem from the other two nodes
> with the same username.  I've looked at show mem, show mem/pool/full, and
> nothing seems to be getting used up.  Is there a sysgen parameter to check? 
> Where else should I look?

I've gotten lot's of responses saying that VIRTUALPAGECNT is probably the
culprit.  I think I need to give more information :-)

I have a cluster with 6610, 6420 and 6410.  They use a common sysdump.dmp and
uaf.  The 6610 and 6420 have idential virtualpagecnt (270144) and memory
(262144 pages).  The 6610 has no problem reading the dump, the 6420 gets
virtual memory exhausted using the same USERNAMES.  They have identical size
pagefiles (and number of pagefiles).

I can try increasing virtualpagecnt, but then I would need to understand why it
would work at the current level on the 6610.  Any further thoughts?

Thanks to all my respondents so far (I tried replying to each, but the numbers
were getting a bit large :-)
-- 
   Don Roberts                   Internet:  don@beckman.com
   Beckman Instruments, Inc.     Phone:     714/961-3029
   2500 Harbor Bl. Mailstop X-12 FAX:       714/961-3351
   Fullerton, CA  92634          Disclaimer:  Blame me, not Beckman.
   So, how many hospitals HAVE been built by organizations of secular
   humanists and atheists, anyway?