• Symptom:
    If your compute node uses a Ceph primary storage, the following error is displayed when you create, enable, or migrate a VM instance:
    Internal error: could not get value of the secret for username 'zstack' using uuid 'ee5b6375-4564-446b-ac53-c94ea4e13904'
  • Cause:

    Generally, when your compute node connects to the Ceph primary storage, the private key for the Ceph primary storage is lost, so you need to delete this private key and reconnect the host.

  • Solution:
    To solve this problem
    1. Run virsh secret-list to obtain the UUID of this Ceph primary storage on the faulty compute node.

      For example: Obtain the 03b5ee0d-df21-46d2-9927-7c930333cb70 UUID.

      root@ceph-host ~]# virsh secret-list 
 UUID                                  Usage
--------------------------------------------------------------------------------
 03b5ee0d-df21-46d2-9927-7c930333cb70  ceph 03b5ee0d-df21-46d2-9927-7c930333cb70
    2. Run virsh secret-undefine 03b5ee0d-df21-46d2-9927-7c930333cb70 to delete the private key of this Ceph primary storage.
      [root@ceph-host ~]# virsh secret-undefine 03b5ee0d-df21-46d2-9927-7c930333cb70
secret 03b5ee0d-df21-46d2-9927-7c930333cb70 deleted
    3. Run /etc/init.d/zstack-kvmagent stop to stop the agent of the compute node.
      [root@ceph-host ~]# /etc/init.d/zstack-kvmagent stop
2017-05-08 10:07:57,590 DEBUG [zstacklib.utils.shell] /sbin/iptables-save
2017-05-08 10:07:57,596 DEBUG [zstacklib.utils.iptables] removed empty chains:[]
2017-05-08 10:07:57,598 DEBUG [zstacklib.utils.shell] /sbin/iptables-restore < /tmp/tmpBW4mWX
Stop Daemon...
Stop Daemon Successfully
stop zstack kvm agent .... SUCCESS
    4. Reconnect this compute node on the UI. After this compute node is reconnected successfully, you can normally create, enable, or migrate VM instances.
      Note: Currently, ZStack has supported the option to disable the CephX private key authentication.