Bug 1038525

Summary: [RFE] [engine-backend] [backup-api] creation of a large number of snapshots as part of the backup api feature without the ability to delete them (live merge), will harm system performance
Product: Red Hat Enterprise Virtualization Manager Reporter: Elad <ebenahar>
Component: ovirt-engineAssignee: Adam Litke <alitke>
Status: CLOSED DUPLICATE QA Contact: Aharon Canan <acanan>
Severity: high Docs Contact:
Priority: unspecified    
Version: 3.3.0CC: abaron, acathrow, amureini, iheim, lpeer, Rhev-m-bugs, scohen, yeylon
Target Milestone: ---Keywords: FutureFeature, Reopened, Triaged
Target Release: 3.5.0   
Hardware: x86_64   
OS: Unspecified   
Whiteboard: storage
Fixed In Version: Doc Type: Enhancement
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2014-02-02 19:55:18 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: Storage RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Bug Depends On: 1032679, 1054219    
Bug Blocks:    

Description Elad 2013-12-05 09:34:08 UTC 
Description of problem:
The new backup api feature, which allows ISV to backup virtual machines (with their disks), uses the snapshots of the VMs as backup points.
The operation of the ISV is to create snapshots on a preconfigured points in time and backup the snapshots disks and OVFs to an external storage. 

Since RHEV does not have the ability to perform live snapshot merge, when we are talking about a normal customer environment, with ~100 VMs, the existing of a large amount of snapshot can be very dangerous to the system. 

On a scenario of a customer that has ~100 VMs and uses a backup software, which will create snapshots to those VMs every day, we can get to a situation of a very large number of volumes which will create and won't be deleted (because we cannot delete snaphsots while VM is running). 

For example:
Lets say that a customer has 100 VMs
Each VM has ~2 disks
Backup software is configured to take snapshot every day:

After 1 month we will have 100*2*30 = 6000 volumes.
In case RAM memory is also taken as part of the snapshot, it will triple the number of the volumes.


Even if customer will shutdown the VMs in order to merge the snapshot, the merge itself will take a relatively long period and will break his continuously operating.
Comment 2 Vered Volansky 2013-12-05 13:15:51 UTC 
What is actually needed here is live snapshot merge capabilities, which is actually an existing RFE, Hence closing duplicate.
Also note that in some cases bz1032679[RFE] implementation should also help this eliviate the problem.

*** This bug has been marked as a duplicate of bug 647386 ***