Re: RH Server puking when RMAN is writing to NFS and NFS goes down
From: Michael Austin <maustin_at_firstdbasource.com>
Date: Thu, 30 Oct 2008 16:48:49 -0500
Message-ID: <lZpOk.6663$Ws1.108@nlpi064.nbdc.sbc.com>
>> On Thu, 30 Oct 2008 08:35:53 -0700 (PDT), sybrandb
>> <sybrandb_at_gmail.com> wrote:
>>
>>> On 30 okt, 14:38, NetComrade <netcomradeNS..._at_bookexchange.net> wrote:
>>>> Anyone seen this? Any clue?
>>> Are we to guess (or to look in a dictionary) what 'puking' means?
>>> Is there salary deducted for every letter you type? It has always been
>>> unclear whether Oracle supported NFS for which version and in which
>>> circumstances.
>>> We don't have any information from you about the affected Oracle
>>> versions, and which NFS version you are using, and with which buffer
>>> settings you mounted it.
>>> If you can't be bothered to provide any detail, please consider to not
>>> post this riddle at all!
>> RH 4
>> Oracle 10.2.0.3
>> nfs mounted with
>> (rw,bg,intr,hard,timeo=600,wsize=32768,rsize=32768,tcp)
>> nfs is used purely for backups
>> by default, it uses ver=3
>> Sun V40Z AMD server
>>
>> By puking, I mean Server dies (reboots).
>> This is less of an oracle issue and more of an OS issue, but I was
>> curious if anyone has seen this.
>>
>> Apologies for not providing details earlier.. was sort of a quick
>> question, maybe someone has seen it.
>>
>> Thanks,
>>
>> .......
>> We run Oracle 9iR2,10gR2, 10g2RAC on RH4/RH5 and Solaris 10 (Sparc)
>> We use RMAN and remote catalog for backups
Date: Thu, 30 Oct 2008 16:48:49 -0500
Message-ID: <lZpOk.6663$Ws1.108@nlpi064.nbdc.sbc.com>
Matthias Hoys wrote:
> "NetComrade" <netcomradeNSPAM_at_bookexchange.net> wrote in message > news:i6njg4h5mr03rd2e6hgvnmteue1oa3eoru_at_4ax.com...
>> On Thu, 30 Oct 2008 08:35:53 -0700 (PDT), sybrandb
>> <sybrandb_at_gmail.com> wrote:
>>
>>> On 30 okt, 14:38, NetComrade <netcomradeNS..._at_bookexchange.net> wrote:
>>>> Anyone seen this? Any clue?
>>> Are we to guess (or to look in a dictionary) what 'puking' means?
>>> Is there salary deducted for every letter you type? It has always been
>>> unclear whether Oracle supported NFS for which version and in which
>>> circumstances.
>>> We don't have any information from you about the affected Oracle
>>> versions, and which NFS version you are using, and with which buffer
>>> settings you mounted it.
>>> If you can't be bothered to provide any detail, please consider to not
>>> post this riddle at all!
>> RH 4
>> Oracle 10.2.0.3
>> nfs mounted with
>> (rw,bg,intr,hard,timeo=600,wsize=32768,rsize=32768,tcp)
>> nfs is used purely for backups
>> by default, it uses ver=3
>> Sun V40Z AMD server
>>
>> By puking, I mean Server dies (reboots).
>> This is less of an oracle issue and more of an OS issue, but I was
>> curious if anyone has seen this.
>>
>> Apologies for not providing details earlier.. was sort of a quick
>> question, maybe someone has seen it.
>>
>> Thanks,
>>
>> .......
>> We run Oracle 9iR2,10gR2, 10g2RAC on RH4/RH5 and Solaris 10 (Sparc)
>> We use RMAN and remote catalog for backups
> > I don't know if this helps, but I saw this in the 10.2.0.4 patch notes: > > "14.16 NFS Client Causes Machine To Reboot During Database Stress Testing > > Under certain test conditions with the system running at about 75% CPU > utilization, a system/OS reboot has been observed on Suse SLES10. An OS bug > has been opened against SLES10 with Novell. The Oracle bug 6883690 contains > the details and references the Novell bug (# 361920) filed for this issue." > > > >
and this is what happens when you use a toy OS to do real work... Received on Thu Oct 30 2008 - 16:48:49 CDT