[pve-devel] KVM Live Backup performance

[Cesar Peschiera]

In a previous post i read (Tue Nov 26 18:45:15 CET 2013):

- Eric Blevins say:
> I have a suggestion that would help alleviate the read and write downsides to
> this.
> 
> Create a memory buffer where the reads/writes from the VM are placed.
> When buffer is over a certain percentage, stop the backup read operations and
> flush the buffer.
> The VM can perform IO up to the limit of the memory buffer with little loss in
> performance.

- And Dietmar Maurer answers:
That is how it works already.

Anyways, I will try to upgrade KVM to 1.7 first (many backup related changes). 
We can then test again and try to optimize further.

Then, I say now:
-----------------------
The mbuffers for writes can be a serious problem if the PVE Host that have the VM in HA is decomposes during the process of backup, because i will lose data of this VM when this VM starts in the other PVE Node that is alive

And my Questions:
---------------------------
1- "KVM live Backup" with the new code of backup, will do write cache for use of the VM?
2- If the question above is correct, how much write cache will make this new code of backup?
And the question more important:
3- If the first question is correct, is possible to make a new code of "KVM live Backup" that don't make write cache for use of the VM and nor affect the I/O performance of the VM?

Best regrads
Cesar

----- Original Message ----- 
From: "Dietmar Maurer" <dietmar at proxmox.com>
To: "Cesar Peschiera" <brain at click.com.py>; <pve-devel at pve.proxmox.com>
Sent: Tuesday, January 28, 2014 2:52 AM
Subject: RE: [pve-devel] KVM Live Backup performance


> If is possible without lose performance into this VM, the write cache for "KVM
> Live Backup" not must to execute it. In this mode the "KVM Live Backup" will be
> fantastic.

Sorry, but I do not really understand that question?

We have done many improvement on backup code, so you should first test with the newest binaries
available on the 'pvetest' repository.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.proxmox.com/pipermail/pve-devel/attachments/20140128/fd63a7d6/attachment.htm>