HACMP failover fails - Aix

This is a discussion on HACMP failover fails - Aix ; Hello all, today one of the two nodes in our HACMP cluster failed. The standby node should have taken over, but clverify failed. This is what I have found in the logs: ""BEGIN CHECK Verifying Fast Disk Takeover Verifying volume ...

+ Reply to Thread
Results 1 to 2 of 2

Thread: HACMP failover fails

  1. HACMP failover fails

    Hello all,

    today one of the two nodes in our HACMP cluster failed.
    The standby node should have taken over, but clverify
    failed. This is what I have found in the logs:

    ""BEGIN CHECK
    Verifying Fast Disk Takeover
    Verifying volume group first_vg for fast disk takeover
    Node node1 reports an LVM level of 503000030 and mode 32 for volume group
    fist_vg
    Node node2 reports an LVM level of 503000030 and mode 32 for volume group
    first_vg
    Volume group first_vg is validated for fast disk takeover
    Verifying volume group second_vg for fast disk takeover
    Node node1 reports an LVM level of 503000030 and mode 32 for volume group
    second_vg
    Node node2 reports an LVM level of 503000030 and mode 0 for volume group
    second_vg
    FAIL

    ERROR: Volume group second_vg is defined as enhanced concurrent
    and the LVM level allows use of fast disk takeover on node
    node1, but not on node node2.
    This inconsistent definition cannot be supported."

    Corrective action failed as well. I varied the vg on and off
    manually on node2 and could start the cluster services through smit.

    All VGs were set to enahnced concurretn durcing HACMP implementation.
    Everything worked during or last test. I have no idea how the definition
    of the vg on node2 could have been changed. No 'chvg' in shell history
    or smit.log on the node. Any other operation that could have caused this?

    Any hints are greatly appreciated.

    Regards,
    uwe

  2. Re: HACMP failover fails

    Uzytkownik "Uwe Weber" napisal w wiadomosci
    news:6ilbvdFredq2U1@mid.individual.net...
    > Hello all,
    >
    > today one of the two nodes in our HACMP cluster failed.
    > The standby node should have taken over, but clverify
    > failed. This is what I have found in the logs:
    >
    > ""BEGIN CHECK
    > Verifying Fast Disk Takeover
    > Verifying volume group first_vg for fast disk takeover
    > Node node1 reports an LVM level of 503000030 and mode 32 for volume group
    > fist_vg
    > Node node2 reports an LVM level of 503000030 and mode 32 for volume group
    > first_vg
    > Volume group first_vg is validated for fast disk takeover
    > Verifying volume group second_vg for fast disk takeover
    > Node node1 reports an LVM level of 503000030 and mode 32 for volume group
    > second_vg
    > Node node2 reports an LVM level of 503000030 and mode 0 for volume group
    > second_vg
    > FAIL
    >
    > ERROR: Volume group second_vg is defined as enhanced concurrent
    > and the LVM level allows use of fast disk takeover on node
    > node1, but not on node node2.
    > This inconsistent definition cannot be supported."
    >
    > Corrective action failed as well. I varied the vg on and off
    > manually on node2 and could start the cluster services through smit.
    >
    > All VGs were set to enahnced concurretn durcing HACMP implementation.
    > Everything worked during or last test. I have no idea how the definition
    > of the vg on node2 could have been changed. No 'chvg' in shell history
    > or smit.log on the node. Any other operation that could have caused this?
    >
    > Any hints are greatly appreciated.


    Are you sure that noone resized one filsystems on activ node and forgot to
    do it using cspock or (importvg -L)



+ Reply to Thread