Setting prio to high since we expect quite a long time before 3.6 is widely adopted and we want to avoid log errors when it's easy to fix. Patches are ready +++ This bug was initially created as a clone of Bug #1284564 +++ Description of problem: Libvirt 1.2.17 (but maybe before) added a new "Crashed" event which VDSM doesn't support yet. In general, VDSM should be forward compatible with future libvirt events. Version-Release number of selected component (if applicable): 4.16.29 How reproducible: 100% Steps to Reproduce: 1. run VM 2. trigger "Crashed" event inside the VM 3. Actual results: VDSM could not handle the event, dumps a stack trace Expected results: VDSM handles the event, and ignores if it haves to. No stacktraces are generated. Additional info: The stracktrace currently VDSM dumps is because VDSM just fails at the start of the handling of the event. There is no loss of functionalty because VDSM will just log the event without doing any further action
after all this seem to happen only on ppc, x86 guests do not seem to use a pvpanic device which triggers this event. Let's not waste time on 3.5 then