Based on bz#1435335, we should add a section in RHV install guide for hosts (both RHEL and RHVH), to mention that if booting from SAN, the boot LUN should have a different config than the rest of the LUNs that are visible to the host and are served for RHV storage domain. Once bz#1435335 is released, we should suggest blacklisting that LUN on vdsm side.
Until bz#1435335 is released, we should suggest in our docs the following info message: ~~~ If the hypervisor is booting from SAN, it is recommended to add a drop-in multipath config file for the boot LUN, as showed in the below example. It should help avoid getting the host filesystem read-only, if something going wrong with the boot LUN. $ cat /etc/multipath/conf.d/host.conf multipaths { multipath { wwid xxxyyy no_path_retry queue } } ~~~ Nir, can you ack please?
This is not a temporary solution until bug 1435335 is resolved, but part of the solution for that bug. Ben, can you confirm that this configuration is enough to enable queuing when all paths have failed? Do you recommend additional configuration for this use case?
This should work to force queuing on a specific multipath device.
Updated document: https://access.redhat.com/documentation/en-us/red_hat_virtualization/4.2/html-single/installation_guide/#Introduction_to_Storage
Test sync