Conference orarep::nomahs::rdb_60

5067.0. "RDMS$$DETACH_DATABASE + DE 6.1-02" by M5::LWILCOX (Chocolate in January!!) Mon Feb 24 1997 12:47

6.1-02 on vax, customer is experiencing bugchecks with accvio and exception
RDMS$$DETACH_DATABASE + DE.  They have run out of pagefile quota according
to the getjpi section in the bugcheck.  They have pagefile quota at 400,000
and virtual page count at 532,288.  She will up the quotas for the moment.

I saw a number of references to this module with different offset and notes
about "memory corruption".  Also noted in release notes for a couple of 6.0
and 6.0A release notes.  Just wondering if this might be a similar problem.

The stack is pretty uninteresting, RDMS$$TOP_DETACH_DATABASE + 129,
then KOD$SETSTK_AND_CONTINUE.

The customer knows that the workload at the plants with this problem has
increased tremendously.

Any other info that would be helpful?  Other suggestions I could make to her?

Thanks,

liz

        Liz,
        
        	Are they using global buffers? is it possible this is the
        old problem where the system gives back to the process the
        pagefile quota it needed to create the buffers and it blows the
        pagfilquota when trying to re-acquire it?
        
        Dave

                       <<< Note 5067.1 by NOVA::BRYDEN >>>

>>        	Are they using global buffers? is it possible this is the
>>        old problem where the system gives back to the process the
>>        pagefile quota it needed to create the buffers and it blows the
>>        pagfilquota when trying to re-acquire it?
        
I'd forgotten about that one!  I'll check on the use of gb.

    Maybe somebody should check the sources...