Re: No records in v$asm_disk

Well, you will have the 2 regular device paths lets call them sd2a and sd2b. Then you'll have the powerpath device, let's call that guy sd2c. When I was having trouble getting ASM to work the first time I set up RAC (2 years ago) there was a note on metalink that specifically said not to use the powerpath device (sd2c in the above example). That fixed the problem I was having. Even when you don't use the powerpath device, if one device path fails it will still use the other path.

Finn

On 11/8/07, William Wagman <wjwagman_at_ucdavis.edu> wrote:
>
> Finn,
>
> I'm not sure what you mean when you say you did not use the powerpath
> device as the disk device. Can you give me a bit more detail please.
>
> Thanks.
>
>
> Bill Wagman
> Univ. of California at Davis
> IET Campus Data Center
> wjwagman_at_ucdavis.edu
> (530) 754-6208
>
>
> ------------------------------
> *From:* oracle-l-bounce_at_freelists.org [mailto:
> oracle-l-bounce_at_freelists.org] *On Behalf Of *Finn Jorgensen
> *Sent:* Thursday, November 08, 2007 12:49 PM
> *To:* moabrivers_at_gmail.com
> *Cc:* oracle-l_at_freelists.org
> *Subject:* Re: No records in v$asm_disk
>
>
> The one thing in your explanation that stood out to me was the part about
> the multipath. When I did this I used EMC powerpath and I specifically DID
> NOT use the powerpath device as the disk device. This caused problems. Even
> when you use the regular device it still uses the multipathing.
>
> Another thing is not to allow ASMLIB to reset permissions on reboots. This
> caused a lot of headaches for me.
>
> Did you run */etc/init.d/oracleasm listdisks* and did that list all the
> disks?
>
> Finally, I would recommend setting up 1 node first as a single node
> cluster and then bringing the second node in after the first is all up and
> running with CRS/ASM/RAC db etc.
>
> Finn
>
>
>
> On 11/8/07, Brian Lucas <moabrivers_at_gmail.com> wrote:
> >
> > All, I've been reading through the forums and metalink looking for
> > this particular problem but haven't been able to find the answer. I
> > have 2 nodes (OEL 5) multipathed to our backend storage. Both nodes
> > are configured with asmlib and the CRS daemons are installed and
> > running great on them. The whole setup is 10g release 2.
> >
> > Prior to disk creation, I set the ORACLEASM_SCANORDER in
> > /etc/sysconfig/oracleasm and added "multipath dm" to it. I even added
> > "sd" to ORACLEASM_SCANEXCLUDE so that all scans happen on the
> > multipath channel.
> >
> > I then create the disk with the following (I have tested this from
> > both node1 and node2).
> >
> > #service oracleasm createdisk TESTDATA /dev/dm-4
> >
> > Scanning the asm disks on both nodes as root successfully shows the
> > TESTDATA asm disk (service oracleasm scandisks followed by service
> > oracleasm listdisks).
> >
> > On node1, querying v$asm_disk shows TESTDATA provisioned and ready to
> > go. Node2 never shows it. I have added ORCL:* to the asm_diskstring
> > and bounced the instance on node2 but to no avail.
> >
> > I have blown away the configuration, zeroed out the underlying
> > storage, and started fresh but I get the same results.
> >
> > Here's the SQL output from node2.
> >
> >
> > SQL> show parameter asm
> >
> > NAME TYPE VALUE
> > ------------------------------------ -----------
> > ------------------------------
> > asm_diskgroups string
> > asm_diskstring string ORCL:*
> > asm_power_limit integer 1
> > SQL> select * from v$asm_disk;
> >
> > no rows selected
> >
> > SQL>
> >
> >
> > If I query /dev/oracleasm/disks, I do see the TESTDATA disk owned by
> > oracle:dba on both nodes.
> >
> > Can anyone point me in a direction I haven't tried?
> >
> > Thanks!
> >
> > Brian
> > --
> > http://www.freelists.org/webpage/oracle-l
> >
> >
> >
>
>

--
http://www.freelists.org/webpage/oracle-l