Re: ASM cannot locate disks
- From: Jeremy Schneider <jeremy.schneider@xxxxxxxxxxxxxx>
- To: dedba@xxxxxxxxxx
- Date: Fri, 03 Aug 2012 07:40:29 -0500
It sounds to me like someone else busted the VTOC on the disks while ASM
was running. Solaris probably didn't re-read the VTOC until the
reboot. It's possible that ASM is still there, starting around the
second block or second MB or something, and just "invisible" without the
VTOC.
You might check the VTOC details on the working REDO disk, assume that
the DATA disks were setup in the same way, then re-label the DATA
disks. I know more about linux than solaris... I've fixed partition
tables lots of times, and I'm pretty sure that it's the same on Solaris
-- updating the VTOC shouldn't do anything at all to the actual data on
the disk. If you're lucky, whoever busted the disk headers didn't mess
up the ASM headers and then the ASM disks will magically re-appear after
fixing the VTOC. However if the ASM headers were also damaged then it's
a much more complicated recovery.
-Jeremy
--
Jeremy Schneider
Chicago
+1 312-725-9249
http://www.ardentperf.com
On 08/03/2012 06:31 AM, De DBA wrote:
> Hi,
>
> This was not the first reboot - I made sure to reboot the entire machine
> after setting up ASM for the first time (it's not yet in production :) ). It
> has been rebooted several times after without ill effect. The sysadmin noted
> that the "invalid disk label" error started appearing some time ago in the
> syslog though. The LUNS were labelled, but format currently reports that the
> disk, though formatted, is not labelled. The SAN was not down at any stage.
>
> The devices are delegated whole to the zone, and not mounted either in the
> global zone or the Oracle zone. There are no errors in any log, apart from
> the ones that I posted :(
>
> The permissions are OK, oracle still owns slice 1 as it was originally
> configured. Perhaps there is something going on with the Hitachi SAN and raw
> partitions? I noted that the "Solaris" presented LUNS feature a whole-disk
> device (c...t...d0) that is missing from the "raw" luns...
>
> Cheers,
> Tony
>
> On 03/08/12 21:03, Mark W. Farnham wrote:
>> Also, was this perhaps the first whole machine reboot since the LUNs were
>> configured? If so, check permissions and that the order of boot sequence did
>> not mount them as owned by root instead of oracle or with the wrong group
>> before oracle tries to mount them correctly for ASM. Those errors would be
>> in the system logs if oracle is trying to mount something that is already
>> mounted. You mentioned that this is a zone, so the overall machine reboot
>> may grab the resources differently than just restarting a zone. Your
>> sysadmin should be helpful with the details about whether this is possibly
>> your problem.
>>
>> -----Original Message-----
>> From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx]
>> On Behalf Of Peter Hitchman
>> Sent: Friday, August 03, 2012 6:35 AM
>> To: oracle-l
>> Subject: Re: ASM cannot locate disks
>>
>> Hi,
>> My experience with problems on Solaris with ASM have either been with device
>> permissions or a failure to partition the LUNs in the first place. Have you
>> checked that nothing odd has happened to the partition tables on the LUNs
>> seen as RAW?
>>
>> Regards
>> Pete
>> --
>> //www.freelists.org/webpage/oracle-l
>>
>>
>> --
>> //www.freelists.org/webpage/oracle-l
>>
>>
>>
>>
> --
> //www.freelists.org/webpage/oracle-l
>
>
>
--
//www.freelists.org/webpage/oracle-l
Other related posts:
