Bug 1985973

Summary:	Remove the abort snapshot behavior
Product:	[oVirt] ovirt-engine	Reporter:	Liran Rotenberg <lrotenbe>
Component:	BLL.Virt	Assignee:	Liran Rotenberg <lrotenbe>
Status:	CLOSED CURRENTRELEASE	QA Contact:	Nisim Simsolo <nsimsolo>
Severity:	medium	Docs Contact:
Priority:	unspecified
Version:	4.4.8.6	CC:	ahadas, bugs, dfodor, nsimsolo
Target Milestone:	ovirt-4.4.9	Keywords:	ZStream
Target Release:	4.4.9	Flags:	pm-rhel: ovirt-4.4+
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:	ovirt-engine-4.4.9-1, vdsm-4.40.90.2	Doc Type:	Bug Fix
Doc Text:	Previously, while executing a snapshot without memory to a VM, it created an abort time for the overall process. Now, we have a new timeout config - `LiveSnapshotFreezeTimeout` which relevant to this flow, making an abort before switching the VM volumes and saving the VM from data inconsistency when freezing the file system.	Story Points:	---
Clone Of:		Environment:
Last Closed:	2021-11-20 08:04:20 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	Virt	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Liran Rotenberg 2021-07-26 12:06:01 UTC

After a discussion, it seems the abortion of snapshot job in VDSM doesn't make much sense. The abort is mostly about hitting timeout while calling libvirt (or doing preparations to that call), an operation that should be very fast; while we mostly hit timing issues calling freeze/thaw operations. In those cases we can't really abort, and if we already passed them we probably should let the snapshot operation to finish.

Therefore, the current decision is to remove the abort mechanism.

Comment 1 Liran Rotenberg 2021-07-27 10:51:39 UTC

In further discussion we saw we have 2 main flows:

1. Snapshot with memory - In such case it makes sense to have a timeout, failing the operation and releasing the VM.
   Even so, we need to think about the timeout (currently 30 minutes by default and configurable in engine-config).
   Also, to consider timeout per snapshot. 

2. Snapshot without memory - In this case we usually desire that the snapshot will be completed.
   In this case we may consider to drop the timeout.

Comment 3 Sandro Bonazzola 2021-11-20 08:04:20 UTC

This bugzilla is included in oVirt 4.4.9 release, published on October 20th 2021.

Since the problem described in this bug report should be resolved in oVirt 4.4.9 release, it has been closed with a resolution of CURRENT RELEASE.

If the solution does not work for you, please open a new bug report.