While others seem to have resolved this issue, sadly, we have not. If I load portal the server abends within a day, but nothing is written to the abend log on the first abend (abend in COMN.NSS version 3.23.04, Sep 9 05 according to the console), the server then does a hard lockup and abends a second time.

If portal is not loaded, the server runs about two weeks and abends for various reasons. The last abend after two weeks was caused by Groupwise quickfinder. We dont have a memory leak, DS, NSS, SERVER, Groupwise memory usage stabilizes (no steady incrementation). Memory fragmentation may be an issue but without Portal it is hard to determine. Setting a static NSS MBCS value (102400) and disabling automatic memory balancing seems to increase the time the server remains stable. NRM memory tuning link does not have any recommendations. The only other clue is a report in NRM of available server processes (250 max) below threshold (25) just prior to the abend.

We are using NSS volumes only, with 2 GB of RAM. Novell ABEND analysis was without results. Suggestions are welcome.