This discussion is archived
4 Replies Latest reply: Oct 25, 2012 7:50 AM by 912834 RSS

Oracle restart starts disabled ASM diskgroup?

912834 Newbie
Currently Being Moderated
I'm confused. I've read the documentation which hasn't helped me. I've scoured the forums and can't find a similar problem, so thought I'd open my question to this forum.

I may have misunderstood, but my understanding is that should a service, instance or diskgroup be disabled on a node, then it should not be started on restart. However upon restart I find a diskgroup which is disabled on a node is started on that node. For info I am running 11gR2 11.2.0.3 x86-64 clusterware on RHEL5.

I have a new installation of RAC and I've been doing some testing. I created several diskgroups including DATA and FRA after creating my ASM instance. I have separate diskgroups to hold the ASM spfile, vote and ocr, so currently DATA and FRA are not being used at all.

I created DATA and FRA on node1. This started them on node1. I then started FRA on node2. So initial state of DATA and FRA:
ora.DATA.dg
               ONLINE  ONLINE       node1
               OFFLINE OFFLINE      node2
ora.FRA.dg
               ONLINE  ONLINE       node1
               ONLINE  ONLINE       node2
I then issued the following:
node1> srvctl stop diskgroup -g FRA -n node1
node1> srvctl disable diskgroup -g FRA -n node1
Both completed successfully. New status:
ora.DATA.dg
               ONLINE  ONLINE       node1
               OFFLINE OFFLINE      node2
ora.FRA.dg
               OFFLINE OFFLINE      node1
               ONLINE  ONLINE       node2

node1> srvctl status diskgroup -g FRA
Disk Group FRA is running on node2
I've performed two separate tests with the same results. test 1 - bounce node1, test 2 - stopped the whole cluster and restarted:
node1> crsctl stop cluster -all
node1> crsctl start cluster -all
The status after restart:
ora.DATA.dg
               ONLINE  ONLINE       node1
               OFFLINE OFFLINE      node2
ora.FRA.dg
               ONLINE  ONLINE       node1
               ONLINE  ONLINE       node2
To me the FRA being ONLINE on node1 is totally unexpected, I would expect that since FRA is disabled on node1, it should not be started. Interestingly DATA which has not yet been started on node2 does not start on node2.

Did the disable command work? I tried it again:
node1> srvctl disable  diskgroup -g FRA -n node1
PRCC-1013 : FRA was already disabled on node1
PRCR-1003 : Resource ora.FRA.dg is already disabled
Does this seem wrong to anyone else?
Can anyone explain the restart behaviour?

Thanks for reading,
Andy
  • 1. Re: Oracle restart starts disabled ASM diskgroup?
    userHH0815 Explorer
    Currently Being Moderated
    Is there an entry in the ASM-spfile "asm_diskgroups=" ?
  • 2. Re: Oracle restart starts disabled ASM diskgroup?
    Sebastian Solbach (DBA Community) Guru
    Currently Being Moderated
    Hi,

    have you created any database yet? Or have you created an ACFS Filesystem?

    Regards
    Sebastian
  • 3. Re: Oracle restart starts disabled ASM diskgroup?
    912834 Newbie
    Currently Being Moderated
    Hi userHH0815,

    Thanks for your reply. I hadn't thought of the ASM SPFILE, so I had a look and there are indeed entries:

    +ASM1.asm_diskgroups='OCR','VOTE','DATA','FRA'#Manual Mount
    +ASM2.asm_diskgroups='OCR','VOTE','FRA'#Manual Mount

    It does seem likely that this is the cause of the FRA starting on both nodes. It is certainly logical that this would cause the diskgroup to start. But at the same time it is surprising that disabling the diskgroup via srvctl has not overruled this.

    I wonder if I'm just trying to do something with Oracle restart that it was never designed to do?


    Also I can confirm there is no databse nor filesystem in DATA or FRA, they are completely empty (for the moment).

    Thanks,
    Andrew
  • 4. Re: Oracle restart starts disabled ASM diskgroup?
    912834 Newbie
    Currently Being Moderated
    I'm even more confused now. I stopped the cluster on node1 only. I then updated the SPFILE...
    alter system set asm_diskgroups=OCR,VOTE,DATA scope=both sid='+ASM1';
    Restarted the cluster on node1 and...
    ora.DATA.dg
                   ONLINE  ONLINE       node1
                   OFFLINE OFFLINE      node2
    ora.FRA.dg
                   ONLINE  OFFLINE      node1
                   ONLINE  ONLINE       node2
    It has certainly kept the diskgroup offline on node1, but not quite the effect I was expecting.

Legend

  • Correct Answers - 10 points
  • Helpful Answers - 5 points