Comment 4 for bug 1036902

Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Fix merged to nova (stable/essex)

Reviewed: https://review.openstack.org/11387
Committed: http://github.com/openstack/nova/commit/09217abddc07bd4fbaca6c300075df2c68ffedf7
Submitter: Jenkins
Branch: stable/essex

commit 09217abddc07bd4fbaca6c300075df2c68ffedf7
Author: Rafi Khardalian <email address hidden>
Date: Mon Aug 13 20:53:43 2012 +0000

    Restore libvirt block storage connections on reboot.

    Fixes bug 1036902 -- Backported version for stable/essex.

    There are a number of cases where block storage connections are not
    properly restored, impacting libvirt in particular. The most common
    case is a VM which has block storage attached via iSCSI, whereby the
    physical system is rebooted. When the system comes back up and
    starts nova-compute, the iSCSI connections are NOT recreated for the
    instances slated to be resumed (assuming
    resume_guests_state_on_host_boot is set).

    The patch changes the compute manager to pass block_storage_info via
    driver.reboot() and driver.resume_state_on_host_boot(). The fix is
    actually only present in the libvirt driver. However, all the other
    drivers were updated to accept the additional, optional function
    arg.

    With the changes in place, iSCSI connections for libvirt are
    re-established after a hypervisor reboot with
    resume_guests_state_on_host_boot=True and on every hard_reboot.
    The latter is intended so that users have a last ditch option for
    recovering their VMs without administrative involvement.

    Change-Id: I8ab3a138b559ee0aa1535a928282e9c372ec5651