Discussions specific to the VMware vSphere hypervisor
Post Reply
jp1
Enthusiast
Posts: 27
Liked: 1 time
Joined: Jun 14, 2013 7:40 pm
Full Name: Jonathan Palsi
Contact:

Processing Rate Question

Post by jp1 » Jun 17, 2013 3:04 pm

Hello all.

I'd like to start off by saying that I'm new to Veeam. In fact, I'm just in the setup and testing stages of bringing up our first virtual environment using ESXi 5.1. We have 3 HP DL3060G8 hosts and an HP P2000 iSCSI SAN w/18 900GB 10K disks. The plan is to use 2 hosts as production servers and the 3rd will run Veeam and be a temporary host if the 2 main ones have issues. Host 3 has direct attached storage w/5 1TB 7.2K disks where we will be saving the Veeam backups to. All three hosts have 2 dedicated NIC's each to the SAN.

As a test, I installed Veeam in a Windows 2008R2 VM on host 3 and created another Windows 2008R2 VM on host 2 w/1 40GB and 150GB disks. I fired off a backup job and got a processing rate of 81 MB/s and the bottleneck listed as Source. Is 81 MB/s good, bad, or it depends? It took approximately 40 minutes to backup the VM.

Thanks!

foggy
Veeam Software
Posts: 18439
Liked: 1588 times
Joined: Jul 11, 2011 10:22 am
Full Name: Alexander Fogelson
Contact:

Re: Processing Rate Question

Post by foggy » Jun 17, 2013 3:22 pm

Jonathan, this is indeed a good speed for the full backup. The bottleneck stats show the percent that every component involved in the backup process is busy so there will always be a bottleneck in every setup. Bottleneck source indicates that the disk reader component spends all of the time reading the data, because the following stages are always ready to accept more data for processing. Please refer to the sticky FAQ topic for further explanation.

yizhar
Service Provider
Posts: 181
Liked: 48 times
Joined: Sep 03, 2012 5:28 am
Full Name: Yizhar Hurwitz
Contact:

Re: Processing Rate Question

Post by yizhar » Jun 17, 2013 11:54 pm

Hi.

Performance looks good but maybe you can get better.
Are you using SAN mode?
Can you post the whole job results here?

Yizhar

jp1
Enthusiast
Posts: 27
Liked: 1 time
Joined: Jun 14, 2013 7:40 pm
Full Name: Jonathan Palsi
Contact:

Re: Processing Rate Question

Post by jp1 » Jun 18, 2013 7:33 pm

Hopefully these are the results you meant, I got these from the Statistics of the completed job.

Duration: 40:03
Processing rate: 81 MB/s
Processed 190GB
Read 186GB
Transferred 5.2GB (35.9x)

VM size: 190GB
Load: Source 99% > Proxy 78% > Network 0% > Target 0%
Primary Bottleneck: Source

Again, the VM being backed up is in a datastore that resides on the SAN. The VM running Veeam is in a datastore that resides on DAS on host 3. The repository is a drive (vdmk) in that VM.

This is my first crack on setting up Veeam and will take any and all suggestions on changes.

Thanks!!

veremin
Product Manager
Posts: 17047
Liked: 1470 times
Joined: Oct 26, 2012 3:28 pm
Full Name: Vladimir Eremin
Contact:

Re: Processing Rate Question

Post by veremin » Jun 19, 2013 8:58 am

It’s clear from the information given that the source disk reader spend most of the time reading the data from the storage, while other backup components are idling.

So, I’m wondering what backup proxy mode is being used. Is it a Direct SAN, Hot-Add, or Network one? This information can be taken from the corresponding job session statistics (look for special [san], [hot-add], [nbd] metrics).

Thanks.

jp1
Enthusiast
Posts: 27
Liked: 1 time
Joined: Jun 14, 2013 7:40 pm
Full Name: Jonathan Palsi
Contact:

Re: Processing Rate Question

Post by jp1 » Jun 19, 2013 2:46 pm

According to the statistics: "Using source proxy VMware Backup Proxy [hotadd].

Vitaliy S.
Product Manager
Posts: 23183
Liked: 1602 times
Joined: Mar 30, 2009 9:13 am
Full Name: Vitaliy Safarov
Contact:

Re: Processing Rate Question

Post by Vitaliy S. » Jun 19, 2013 10:18 pm

Then you're already getting maximum from your storage, as HotAdd is the second most optimal processing mode. Configuring direct SAN backup mode might give you slightly better job performance rates, but the difference will not be noticeable during incremental job runs.

jp1
Enthusiast
Posts: 27
Liked: 1 time
Joined: Jun 14, 2013 7:40 pm
Full Name: Jonathan Palsi
Contact:

Re: Processing Rate Question

Post by jp1 » Jun 20, 2013 11:28 am

Thanks guys.

I did add a bunch of files to the test VM and ran an incremental. I believe that produced a rate of something 170 MB/s, so it sounds like I'm on the right track. Unfortunately I can't give exact stats since I deleted the VM running Veeam since it was only running as a test and now I'm setting everything up for production.

foggy
Veeam Software
Posts: 18439
Liked: 1588 times
Joined: Jul 11, 2011 10:22 am
Full Name: Alexander Fogelson
Contact:

Re: Processing Rate Question

Post by foggy » Jun 20, 2013 11:34 am

jp1 wrote:I did add a bunch of files to the test VM and ran an incremental. I believe that produced a rate of something 170 MB/s, so it sounds like I'm on the right track.
Note that incremental runs are typically a few times faster than the full backup.

Post Reply

Who is online

Users browsing this forum: No registered users and 10 guests