FW: Solaris and Checkpoint

[Ray Zupancic <rzupancic () corp webb net>]

I'm new to the list, and assume I can submit general product/hardware
configuration problems.   

I inherited a network with a Checkpoint 4.0 FW-1 on an Ultra5 running
Solaris 2.6 with 128 megs of ram.  Originally it was two identical systems
clustered with Veritas Firstwatch

In any case, this system mysteriously slows to a crawl and stops processing
traffic at random times of the day--usually up to 10 minutes or so--and then
resumes without intervention.  The only thing I can see is that the system
usage goes to 100, and the run queue skyrockets due to this (there is not
process running away, it seems to be kernel).

I've shutdown everything I can including the Veritas to try to run this
problem down--no luck so far.  I was told the system could process up to
20Mbs, and we have less than 4Mbs.

If the front-end web server load is taken off, it runs pretty much perfectly
(much less load).  If anyone has any ideas or theories on this problem,
please let me know.  Support hasn't been very useful on this issue.


Thanks,

--rayz