Re: Physical disk hangs at "offline pending"

Hi,

You are probably right about the group configuration and I will do some
research into it today.

Since the group move works fine from node 1 to node 2 but doesn't work going
from node 2 to node 1 I'm not sure that group configuration is causing my
problems.

Even with all the oracle services offline and only the Cluster IP, name,
quorum and the 1 physical disk online a move from node 2 to node 1 leaves the
physical disk as "offline pending" while a move from node 1 to node 2 works
fine.

I am going to seeif I can find any differences that are not obvious between
the nodes first. Then I will try re-arranging the groups.

Any other suggestions would be appreciated.

Thanks in Advance,
--
Henry


"Edwin vMierlo" wrote:

Henry,

I think here we go wrong.

The disks your Oracle instances are using are not in the same group of your
Oracle instances.

so, this is what I think here is happening.
Cluster group online on node 1
Oracle group online on node 1

Oracle is happy, as its disks are accessible.... as they are all on node 1.

Cluster group wants to move over, tries to offline the disks, however your
Oracle instances still have a handle open on the disks, and therefore the
disk remains off-line pending until time out.

I seriously believe you need to change your configuration.

The cluster group should only contain the following :
1) Quorum disk
2) Cluster IP address
3) Cluster Network Name

Each of your Oracle groups need the following
1) IP address
2) Network Name
3) as many disks as this Oracle instance need, these disks can only be used
by this instance exclusively
4) Oracle services / applications for this instance.

Other than this, follow Oracle instructions for building your cluster,
invoke Oracle support for getting more information on how to configure your
clustered applications.

Rgds,
Edwin.


"Henry" <Henry@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:DEE12401-3D9C-4879-8CEC-792EB43A6BBA@xxxxxxxxxxxxxxxx
Hi,

Pardon my ignorance but I'm not sure what is meant by "take offline
everything but the cluster". If that means the cluster group, it was
modified
and contains:
Cluster IP
Cluster Name
Quorum disk physical disk
Oracle log physical disk
Oracle DB physical disk
Images physical disk
Oracle Services for MSCS
Several file shares dependant on the physical disks.

The other groups (2 for different instances) contain:
Oracle Database
Oracle failover IP address
Oracle Listener
Oracle network name

This is not how I originally configured it but it was changed by the DBA
that installed Oracle failsafe.

The two servers are identical (HW and OS).

Thanks in Advance,
--
Henry


"kkantchev@xxxxxxxxxxx" wrote:

Hi,
In order to determine who is holding the HD resource, do the following.
1. Restart and bring everything up and running on node 1.
2. Take offline everything except the cluster.
3. Do failover to Node 2
4. Bring online only the HD in questin
5. Do failover back to Node1

If step 5 is not succeseful, then there is process external to the
cluster
that i causing the problem. Antivirus scan or something like that.
If step 5 is succeseful, then the problem is in your Oracle setup. In
this
case you have to send more details about your configuration.

Hope that helps

Krass



.

Relevant Pages

  • Re: Second HBA to east io on E250?
    ... >> We have one of these venerable old machines hosting an Oracle application ... >> and a message queue on it's internal disks. ... >> with anything made by Oracle plus a message queue. ...
    (comp.sys.sun.admin)
  • Re: I/O performance troubleshooting
    ... maybe oracle hasnt enough sga to optimize what it sends to disk. ... but your poor memory-usage could be improved by increasing read-ahead. ... It seems pretty obvious to me my performance problem is with the disks I have. ... r b avm fre re pi po fr sr cy in sy cs us sy id wa hr mi se ...
    (AIX-L)
  • [HPADM] RE: Oracle Database on separate physical disks
    ... redo logs are put on separate physical disks (because of the sequential I/O ... If you have an Oracle DBA on site, ask him to find out the database tables ...
    (HP-UX-Admin)
  • Re: Physical disk hangs at "offline pending"
    ... The disks your Oracle instances are using are not in the same group of your ... as its disks are accessible.... ... Cluster group wants to move over, tries to offline the disks, however your ...
    (microsoft.public.windows.server.clustering)
  • Daily Report #4165
    ... The resultant cleaned cluster CMDs will ... well-understood host galaxy environment. ... The Nature of the Halos and Thick Disks of Spiral Galaxies ... ACS, NICMOS, and WFPC2 in parallel. ...
    (sci.astro.hubble)