Knowledge Base  /  Disaster Recovery  /  SafeHaven 3
Knowledge Base  /  Disaster Recovery  /  SafeHaven 3

SafeHaven: Gracefully Shutdown SRN Node

Updated by Jake Malmad on Oct 29, 2015
Article Code: kb/620


Rebooting an SRN without taking the proper steps for a graceful shutdown can cause problems with the protection groups residing on the SRN. The most common is reconstitution failure, which renders protection groups in an inactive or error state. This guide walks you through the steps required for a proper SRN shutdown. Note that stopping/restarting services does not have the same potential and ,as such, can be performed at any time.


  1. Login to SRN and kill the voffice service with the Killall -9 voffice command.

  2. Login to the guest vm and disconnect from the iSCSI target:

    • Windows: Disconnect from Iscsi target from Iscsi initiator.
    • Linux: iscsiadm –m node –u
  3. Login to SRN via SSH.

  4. Remove the Iscsi targets:

    • scstadmin -list_target
    • Perform the following command with each IQN pasted into the field: scstadmin -disable_target <the target name> -driver iscsi
  5. Deactivate the Pluns (logical volumes) for the protection groups:

    • Run lvscan.
    • Run vgchange –an <VG name> using the volume names, for each volume.
    • Run lvscan again to make sure all volumes have been deactivated (lvscan should show all as "inactive"). If you are unable to deactivate a volume, run kpartx -d /dev/sbd0. Then kpartx -d /dev/sbd1 (use all sbd's from 0 onward).
  6. Clear all the sbd’s : sbdAllClear (or individually clear sbd's: sbdm /dev/sbd<number> clear similar to part 3 of Step 5).

  7. Clear the bitmaps with sbdFlushAllBitmaps.

  8. Reboot the SRN.

  9. After rebooting the SRN, make sure that the Protection Groups are reconstituted and in "running state" and that the Guest connects to the targets again. On the Production (source) VM, run DgSyncEx.exe list from the SafeHaven Tools Command Prompt to make sure sync is intact. Check lvscan to make sure that the logical volumes are active, and run scstadmin –list_target to verify that the iSCSI targets are being presented.

  10. (Optional) Create a new checkpoint.