On May 10, 6:53 pm, NetComrade <netcomradeNS..._at_bookexchange.net>
wrote:
> On 10 May 2007 10:42:54 -0700, Valentin Minzatu
>
>
>
>
>
> <valentinminz..._at_yahoo.com> wrote:
> >On May 10, 1:11 pm, NetComrade <netcomradeNS..._at_bookexchange.net>
> >wrote:
> >> I looked at profiles, and you can only limit CPU/reads and 'private
> >> sga', which only applies to MTS.
>
> >> We have a customer whose database crashes the whole box (kernel
> >> panic). They think it's due to memory overconsumption. None of the sar
> >> logs, or even 'pga snapshots' that we take indicate that.
>
> >> Unfortunately, they're a bit reluctant to upgrade to the latest redhat
> >> version (they're on RH3U2). We want them to get to at least the latest
> >> redhat update on RH3 so that we could get crash dumps.
>
> >"database crashes the whole box " sounds like an H/W or O/S problem to
> >me. I have experienced similar behavior on a RH3U3 box. Eventually the
> >sys admin discovered that one of the HBAs was at fault.
>
> Unfortunately this happens on both (identical) boxes. There is nothing
> in the logs to indicate any issues. The box crashes with a kernel
> panic. Up until recently, the frequency of the crashes was about every
> 2 weeks (notice about, it wasn't a particular day, or a particular
> time that stayed out). Even if we 'failed over' to a different box,
> the crash would still happen in about two weeks.
> .......
> We run Oracle 9iR2,10gR1/2 on RH4/RH3 and Solaris 10 (Sparc)
> remove NSPAM to email- Hide quoted text -
>
> - Show quoted text -
In our case it was a little bit "more" specific - the crash would
occur shortly after the rman backup or towards the end of it. *BUT*,
if we ran the backup manually no issue was encountered ...
Received on Fri May 11 2007 - 10:51:35 CDT
