2 Replies Latest reply: Oct 3, 2012 11:26 AM by Valentin Minzatu RSS

    unable mount diskgroup using new disk truecopy

    moracle
      so i have this instance DataGuard Standby using ASM disk
      i want to copy the old asm disk that at USPVM storage to new disk at VSP storage
      im using truecopy
      when i complete a pair, do a split, shutdown instance, dismount old asm diskgroup, change owner (old disk to bin, new disk to oracle owner)
      then i try to mount the diskgroup but it keep throwing me error
      ORA-15066: offlining disk may result in a data loss
      i checked asm alert log and give me this information
      SQL> alter diskgroup DRCDATA mount
      Mon Oct 01 18:03:13 TST 2012
      NOTE: cache registered group DRCDATA number=2 incarn=0x0625874b
      Mon Oct 01 18:03:13 TST 2012
      NOTE: Hbeat: instance first (grp 2)
      Mon Oct 01 18:03:13 TST 2012
      Errors in file /opt/oracle/admin/+ASM/bdump/+asm_gmon_1928.trc:
      ORA-27091: unable to queue I/O
      ORA-27072: File I/O error
      HPUX-ia64 Error: 13: Permission denied
      Additional information: 4
      Additional information: 2044
      Additional information: -1
      Mon Oct 01 18:03:13 TST 2012
      ERROR: too many offline disks in PST (grp 2)
      WARNING: Disk 0 in group 2 in mode: 0x7,state: 0x2 will be taken offline
      Mon Oct 01 18:03:13 TST 2012
      ERROR: PST could not set any hearbeat (grp 2)
      Mon Oct 01 18:03:13 TST 2012
      NOTE: cache dismounting group 2/0x0625874B (DRCDATA)
      NOTE: dbwr not being msg'd to dismount
      ERROR: diskgroup DRCDATA was not mounted
      Mon Oct 01 18:03:13 TST 2012
      Errors in file /opt/oracle/admin/+ASM/bdump/+asm_n000_1947.trc:
      ORA-15001: diskgroup "DRCDATA" does not exist or is not mounted

      i give all the permission that disk needed, any ideas?
      thanks before
        • 1. Re: unable mount diskgroup using new disk truecopy
          onedbguru
          moracle wrote:
          so i have this instance DataGuard Standby using ASM disk
          i want to copy the old asm disk that at USPVM storage to new disk at VSP storage
          im using truecopy
          when i complete a pair, do a split, shutdown instance, dismount old asm diskgroup, change owner (old disk to bin, new disk to oracle owner)
          then i try to mount the diskgroup but it keep throwing me error
          ORA-15066: offlining disk may result in a data loss
          i checked asm alert log and give me this information
          SQL> alter diskgroup DRCDATA mount
          Mon Oct 01 18:03:13 TST 2012
          NOTE: cache registered group DRCDATA number=2 incarn=0x0625874b
          Mon Oct 01 18:03:13 TST 2012
          NOTE: Hbeat: instance first (grp 2)
          Mon Oct 01 18:03:13 TST 2012
          Errors in file /opt/oracle/admin/+ASM/bdump/+asm_gmon_1928.trc:
          ORA-27091: unable to queue I/O
          ORA-27072: File I/O error
          HPUX-ia64 Error: 13: Permission denied
          Additional information: 4
          Additional information: 2044
          Additional information: -1
          Mon Oct 01 18:03:13 TST 2012
          ERROR: too many offline disks in PST (grp 2)
          WARNING: Disk 0 in group 2 in mode: 0x7,state: 0x2 will be taken offline
          Mon Oct 01 18:03:13 TST 2012
          ERROR: PST could not set any hearbeat (grp 2)
          Mon Oct 01 18:03:13 TST 2012
          NOTE: cache dismounting group 2/0x0625874B (DRCDATA)
          NOTE: dbwr not being msg'd to dismount
          ERROR: diskgroup DRCDATA was not mounted
          Mon Oct 01 18:03:13 TST 2012
          Errors in file /opt/oracle/admin/+ASM/bdump/+asm_n000_1947.trc:
          ORA-15001: diskgroup "DRCDATA" does not exist or is not mounted

          i give all the permission that disk needed, any ideas?
          thanks before
          So, apparently truecopy isn't so true? There was a way that you could accomplish this without taking everything offline.

          1) before you ever try to add any new ASM device(s), make sure that ALL nodes can READ AND WRITE to these new devices.
          2) execute the following commands:

          alter diskgroup <DG> drop disk '<path-to-device1>' rebalance power 0;
          alter diskgroup <DG> drop disk '<path-to-devicen>' rebalance power 0;
          alter diskgroup <DG> drop disk '<path-to-devicen>' rebalance power 0;
          alter diskgroup <DG> drop disk '<path-to-devicen>' rebalance power 0;
          ... for ALL devices
          alter diskgroup <DG> add disk '<path-to-new-device1>' rebalance power 0;
          alter diskgroup <DG> add disk '<path-to-new-devicen>' rebalance power 0;
          alter diskgroup <DG> add disk '<path-to-new-devicen>' rebalance power 0;
          alter diskgroup <DG> add disk '<path-to-new-devicen>' rebalance power 0;
          .... for ALL new devices

          then
          alter diskgroup <DG> rebalance power 5;

          Set the power depending on your hardware capabilities. - an HP server could more than likely handle the whole thing. Linux boxes may not be able to handle all of the devices at once, but they can be "replace" on or two at a time. Using this procedure, your system will NOT require any downtime (unless your sys admin cannot add the devices dynamically.
          • 2. Re: unable mount diskgroup using new disk truecopy
            Valentin Minzatu
            It looks like you may have a permission issue (not sure it is the only one though):
            Errors in file /opt/oracle/admin/+ASM/bdump/+asm_gmon_1928.trc:
            ORA-27091: unable to queue I/O
            ORA-27072: File I/O error
            HPUX-ia64 Error: 13: Permission denied
            What does the trace file /opt/oracle/admin/+ASM/bdump/+asm_gmon_1928.trc show?