Description of problem: iSCSI sessions are disconnected early when deleting all volume backed instances from a host. This is due to the instances being shutdown prior to a lookup being preformed of block devices being used by running instances during each disconnect_volume call. This becomes an issue if dm-multipath is used on the host and is configured with queue_if_no_path as systemd-udev etc may hang forever waiting for the iSCSI sessions to reappear. Version-Release number of selected component (if applicable): OSP 7 How reproducible: Always Steps to Reproduce: 1. Run multiple volume backed instances on a single compute node. 2. Remove all of these instances from the host at the same time. Actual results: iSCSI sessions are disconnected early prior to all volumes being disconnected. Expected results: iSCSI sessions are only disconnected during the final disconnect_volume call. Additional info: More details in the following comments where this was initially discovered. https://bugzilla.redhat.com/show_bug.cgi?id=1351169#c55 https://bugzilla.redhat.com/show_bug.cgi?id=1351169#c57
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHBA-2016-2025.html