Francesco Romani has posted comments on this change.

Change subject: recovery: try to restore VMs from recovery files
......................................................................


Patch Set 9: -Verified

(5 comments)

http://gerrit.ovirt.org/#/c/25276/9/tests/functional/virtTests.py
File tests/functional/virtTests.py:

Line 192:             self.assertEquals(result['status'], vmstatus.UP)
Line 193: 
Line 194:     def _waitForBoot(self, vmid):
Line 195:         self.retryAssert(partial(self.assertQemuSetupComplete, vmid),
Line 196:                          timeout=10)
> since we retry I wouldn't be afraid to increase these to something working 
OK, will fix in a future patch.
Line 197:         self.retryAssert(partial(self.assertVmBooting, vmid),
Line 198:                          timeout=3)
Line 199:         self.retryAssert(partial(self.assertVmUp, vmid),
Line 200:                          timeout=10)


http://gerrit.ovirt.org/#/c/25276/9/tests/functional/vmRecoveryTests.py
File tests/functional/vmRecoveryTests.py:

Line 33: 
Line 34: from utils import SUCCESS
Line 35: 
Line 36: 
Line 37: def vdsm_start(wait_secs=5):
> where did we ensure that?
Dan is referring to ensure_vdsm_started (line 101 below). And he's right, this 
is  probably (-> 99%) became redundant. I'll verify and I'll remove if this is 
confirmed.

About startup: I posted a patch for this as well: 
http://gerrit.ovirt.org/#/c/26638/ (seems to fix all the issues locally, needs 
to be tested on jenkins, however)
Line 38:     service_start('vdsmd')
Line 39:     start = time.time()
Line 40:     while time.time() - start < wait_secs and service_status('vdsmd'):
Line 41:         time.sleep(1)


Line 88:         vms = []
Line 89:         for i in range(1, num + 1):
Line 90:             vm = VMProxy(self.vdsm, customizer(i))
Line 91:             vm.start()
Line 92:             self._waitForBoot(vm.id)
> The test could be a bit quicker if you run both vms first, and wait for bot
Done
Line 93:             vms.append(vm)
Line 94: 
Line 95:         try:
Line 96:             yield vms


Line 113:         # #1 disappears:
Line 114:         #    it goes down while VDSM is down: must be reported as 
down.
Line 115:         with self.running_vms(2, customize_vm) as vms:
Line 116:             os.kill(vms[0].pid, signal.SIGTERM)
Line 117:             time.sleep(1)  # just to make sure
> "To be sure" you must waitForDown
Will fix
Line 118: 
Line 119:             stats_before = vms[0].stats()
Line 120: 
Line 121:             pid = vms[1].pid


http://gerrit.ovirt.org/#/c/25276/9/vdsm/clientIF.py
File vdsm/clientIF.py:

Line 425:                                  ' This should not happen!'
Line 426:                                  ' Will try to recover them.', 
len(recVms))
Line 427:             for vmId in recVms:
Line 428:                 if not self._recoverVm(vmId):
Line 429:                     self.log.warning('VM %s failed to recover from 
recovery'
> just to doublecheck - VdsUpdateRuntimeInfo does handle correctly when it ha
Good point. I'll make sure before to tick 'verified' again.
Line 430:                                      ' file, reported as Down', vmId)
Line 431: 
Line 432:             while (self._enabled and
Line 433:                    vmstatus.WAIT_FOR_LAUNCH in [v.lastStatus for v in


-- 
To view, visit http://gerrit.ovirt.org/25276
To unsubscribe, visit http://gerrit.ovirt.org/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: Id495f6047ba658c2b04da19bd7bf76425b3b9659
Gerrit-PatchSet: 9
Gerrit-Project: vdsm
Gerrit-Branch: master
Gerrit-Owner: Francesco Romani <[email protected]>
Gerrit-Reviewer: Dan Kenigsberg <[email protected]>
Gerrit-Reviewer: Francesco Romani <[email protected]>
Gerrit-Reviewer: Michal Skrivanek <[email protected]>
Gerrit-Reviewer: Vinzenz Feenstra <[email protected]>
Gerrit-Reviewer: [email protected]
Gerrit-Reviewer: oVirt Jenkins CI Server
Gerrit-HasComments: Yes
_______________________________________________
vdsm-patches mailing list
[email protected]
https://lists.fedorahosted.org/mailman/listinfo/vdsm-patches

Reply via email to