AW: [suse-oracle] Novell - it is VERY SERIOUS! Weshouldadviseagainst using S 2007-01-09 - By Alexei_Roudnev
Back
Ok, it's something.
Questions to all, who have this bug and who do not have it:
- SAN system/driver?
- MPIO software/mode?
- DO you use partitions? (it is strange question, I know, but I posted/SuSe
fixed one bug, related to unpartitioned devices)
Who can reproduce a bug - can you show exact configuration? I mean -
platform - kernel - lsmod output - mpio if any - FC cards if any...
If you use MPIO - what happen without it (I mean - not turning it down, but
not using md /or other/ driver, normally running with mpio)?
PS. Looks as SLES version and file system are not important for this bug.
Bug is somewhere below, in the disk access. Can it be a bug in SCSI or HBA
driver?
-- -- Original Message -- --
From: "Eduardo D Piovesam" <eduardo@(protected)>
To: "suse-oracle" <suse-oracle@(protected)>
Sent: Tuesday, January 09, 2007 4:20 PM
Subject: Re: AW: [suse-oracle] Novell - it is VERY SERIOUS!
Weshouldadviseagainst using SuSe and Oracle if
Certainlly it's a bug with an specific config.
We use many Oracle databases with AIO + 10.1.0.4 + SLES9SP1_x86 (32
bits) + reiserFS + specific patch for SLES9 and AIO with no issues.
All databases run 24x7, and some with very intensive I/O almost all the
time. No problem so far... running for 200+ days and counting.
We're very "happy" with this config, we use it in almost all our Oracle
installations.
Regards,
Eduardo
On Tue, 2007-01-09 at 13:04 -0800, Alexei_Roudnev wrote:
> You may try to get a statistics, WHO and in WHICH CONFIGURATIONS had this
> bug. For example, I never saw it in the lab, and our data center in Canada
> never saw it in production, but we all use iSCSI and ext3/ASM (and don't
use
> multipath software because iSCSI have embedded reliability).
>
> I don't think, that this problem exists in all configurations - if so,
they
> in SuSe was able to find it many month ago. On the other hand,
> you have a customer, have exact configuration with exact description - so
> it's a matter of time to reproduce it.
>
> What's possible here, as a factor, raising this bug:
>
> - multipath IO
> - specific file system
> - specific IO scheduler
> - specific kernel parameters (unlikely, but who knows)
> - specific response, coming from SAN storage under special conditions (so
> that oter SAN storage wil not show this bug).
>
> I am testing few new servers now, and I will try to get this bug running
> TPCC tests, BUT the chances to have it are low.
>
>
>
> -- -- Original Message -- --
> From: "Arun Singh" <Arun.Singh@(protected)>
> To: "Frank Westheider" <frank.westheider@(protected)>
> Cc: "'Anton Dischner'" <Anton.Dischner@(protected)>;
> <suse-oracle@(protected)>
> Sent: Tuesday, January 09, 2007 11:41 AM
> Subject: Re: AW: [suse-oracle] Novell - it is VERY SERIOUS! Weshouldadvise
> against using SuSe and Oracle if
>
>
> Hi Frank,
>
> I can understand the frustration here as I am following this in the
> bugzilla. You are right in asserting the complexity of the bug.
>
> Regards,
> Arun
>
> >>> On 1/9/2007 at 11:11 AM, "Frank Westheider"
> <frank.westheider@(protected)>
> wrote:
> > Hi Arun,
> >
> > Sorry...but cool down is not an option.
> >
> > We have this bug on 3 machines since beginning of september (!!!) and
> opened
> > novell and oracle support-cases....without any success/effort.
> >
> > We escalated these cases at novell and got 2 kernel-ptf-fixes...we
> tested
> > both and diffed the ptf-kernel-source with the original buggy
> > kernel-source...and even we saw, that the difference were not
> concerning the
> > aio-kernel-part.
> >
> > We can reproduce the problem (and wrote this to the support too) with
> heavy
> > async-io-load over a duration of several days (batch-processes). So i
> think,
> > this would be possible to Novell/Oracle/the kernel-guys as well with
> simple
> > db-jobs using aio and firing until the aio-bug is shown....
> >
> > Head-shaking about the support...beginning of september and no really
> effort
> > to fix this thing. I hope you know, what aio is doing for really
> busy
> > databases...
> >
> > Frank
> >
> > -- --Urspr??ngliche Nachricht-- --
> > Von: Arun Singh [mailto:Arun.Singh@(protected)]
> > Gesendet: Dienstag, 9. Januar 2007 16:34
> > An: suse-oracle@(protected)
> > Betreff: RE: [suse-oracle] Novell - it is VERY SERIOUS! We
> shouldadvise
> > against using SuSe and Oracle if suc
> >
> > Cool down folks. Do any of you have any easy way to trigger this
> kernel bug?
> > I can promise fix will take no time, If we can find a way to simulate
> this
> > bug internally. SUSE support/kernel team is aware of the seriousness
> of this
> > bug.
> >
> > -Arun
> >
> >
> >
> >>>> On 1/9/2007 at 1:16 AM, "Yoav Givon" <YGivon@(protected)> wrote:
> >> I strongly support you Alexei on this one. This is not very
> business
> >> like and I believe Novell Officials should make some strong
> commitment
> >> here other wise they will loose us . Loud and Clear.
> >>
> >> Yoav Givon
> >>
> >> -- --Original Message-- --
> >> From: Alexei_Roudnev [mailto:Alexei_Roudnev@(protected)]
> >> Sent: Monday, January 08, 2007 10:40 PM
> >> To: Anton Dischner; Didier Boiteux
> >> Cc: SuSE MailingList
> >> Subject: [suse-oracle] Novell - it is VERY SERIOUS! We should
> advise
> >> against using SuSe and Oracle if such bugs are not fixed PROMPTLY!!
> Re:
> >> AW: [suse-oracle] kernel bug
> >>
> >> This is a critical bug - why it takes so long to fix it?
> >>
> >> Novell is not primary Linux in Oracle (it is not tested with it
> >> automatically by internal users, as they do with RHEL4.4), so it
> makes
> >> a very important to fix such problems ASAP when they arise!)
> >>
> >>
> >> -- -- Original Message -- --
> >> From: "Anton Dischner" <Anton.Dischner@(protected)>
> >> To: "Didier Boiteux" <dboiteux@(protected)>
> >> Cc: "SuSE MailingList" <suse-oracle@(protected)>
> >> Sent: Monday, January 08, 2007 12:54 AM
> >> Subject: Re: AW: [suse-oracle] kernel bug
> >>
> >>
> >>> Hi,
> >>>
> >>> we had a kernel crash with a system which had the last kernel
> patches
> >>> installed.
> >>>
> >>> Kernel 2.6.5-7.283 did -not- fix the kernel BUG at fs/aio.c:733
> >>> with bug number 165140
> >>>
> >>> regards,
> >>>
> >>> Toni
> >>>
> >>>
> >>> --
> >>>
> +-- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- --+
> >>> | Anton Dischner, Unix admin, Oracle-DBA Phone: +49 89
> 70953202|
> >>> | Institut fuer Klinische Chemie Fax : +49 89
> 70958888|
> >>> | Klinikum Grosshadern Home : +49 89
> 69708766|
> >>> | Ludwig Maximilians Universitaet Muenchen Mobil: +49 172
> 8388880|
> >>> | 81366 Muenchen Germany
> |
> >>> | Marchioninistr. 15 Anton.Dischner"at"med.uni-muenchen.de
> |
> >>>
> +-- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- --+
> >>>
> >>>
> >>>
> >>> --
> >>> To unsubscribe, email: suse-oracle-unsubscribe@(protected)
> >>> For additional commands, email: suse-oracle-help@(protected)
> >>> Please see http://www.suse.com/oracle/ before posting
> >>>
> >>>
> >>
> >
> >
> >
> >
>
>
>
>
> --
> To unsubscribe, email: suse-oracle-unsubscribe@(protected)
> For additional commands, email: suse-oracle-help@(protected)
> Please see http://www.suse.com/oracle/ before posting
>
>
>
--
To unsubscribe, email: suse-oracle-unsubscribe@(protected)
For additional commands, email: suse-oracle-help@(protected)
Please see http://www.suse.com/oracle/ before posting
--
To unsubscribe, email: suse-oracle-unsubscribe@(protected)
For additional commands, email: suse-oracle-help@(protected)
Please see http://www.suse.com/oracle/ before posting
|
|
