Description of problem: BZ907997 introduced Run Once event messages but there's little chaos if it is combined with Paused state. a/ normally run once - stateless (FYI I opened BZ969440 so 2nd msg is in simple past) 1. Starting VM YYYYYY was initiated by admin@internal. (code=USER_INITIATED_RUN_VM) 2. VM YYYYYY was restarted on Host XXXX.redhat.com (code=VDS_INITIATED_RUN_VM) b/ run once - stateless + paused 1. Starting VM YYYYYY was initiated by admin@internal. (code=USER_INITIATED_RUN_VM) 2. VM YYYYYY was resumed by admin@internal (Host: xxxxx). 3. VM YYYYYY was resumed by admin@internal (Host: xxxxx). See '2.' is (probably) about starting the VM from paused mode. Shouldn't then '3.' have '(was) restarted' message to tell an user that the VM is finally up? '/b' applies to Run Once - paused (stateful) as well. Wouldn't be nice to have it like this? * was started/was initiated * was paused (if applicable) * was resumed (if applicable) * started/restarted Anyway, I still don't understand why you didn't make it obvious in the first message like 'was initiated in paused mode' and you have implemented these events in even more cryptic way :-) Version-Release number of selected component (if applicable): sf17.1 How reproducible: 100% Steps to Reproduce: 1. run once - paused (stateful) -> see events 2. run once - paused + stateless -> see events 3. compare with same actions but not in 'paused' mode Actual results: run once - paused mode event messages look odd Expected results: (as whole events are cryptic anyway) still make them more clean to users Additional info:
what's the actual change needed to fix this?
a different messages at different point, not particularly difficult. Just need to try all cases and verify it fires the right events
this bug won't fit into 3.5 release and is being deferred to a later release. If you deeply care about this bug and deserves to be re-evaluated please let me know
fixed long ago, on 3.6 i get (order is from new to old, as shown in the ui): VM lightVm started on Host dubi as stateless VM lightVm was resumed by admin@internal (Host: dubi). VM lightVm has been paused. Starting VM lightVm was initiated by admin@internal. please verify
Verified with rhevm-3.6.0.3-0.1.el6.noarch. 1. runOnce Vm as stateless + pause mode: "Starting VM YYYYY was initiated by admin@internal." "VM YYYYY has been paused." after resuming VM: "VM YYYYY was resumed by admin@internal (Host: host_mixed_3)." "VM YYYYY started on Host host_mixed_3 as stateless" 2.runOnce Vm as statefull + pause mode: "Starting in paused mode VM YYYYY was initiated by admin@internal." "VM YYYYY has been paused." after resuming VM: "VM YYYYY was resumed by admin@internal (Host: host_mixed_3)." "VM YYYYY started on Host host_mixed_3" 3.runOnce Vm as stateless normally (not paused): "Starting VM YYYYY was initiated by admin@internal." "VM YYYYY was restarted on Host host_mixed_3 as stateless This is almost perfect and much better than results in the bug description, with the small exception that I would expect that in case (1) the initial message would be: "Starting in paused mode VM YYYYY was initiated by admin@internal." similar to case (2). The behaviour is valid and good and so I am verifying and adding needinfo to raise the question whether we want to make this last adjustment I mentioned or not.
I agree and the fix is trivial (https://gerrit.ovirt.org/#/c/48840/1). If you think it should be backported to 3.6 then reopen the bug. I think that since it is not part of this bug and the severity is low, we can push it to 4.0. in that case, worth opening a bug for 4.0 so it will be verified then.