Expected speed for LAN NAS backup

[spgsit5upport]

Ofcourse nobody will give 100% correct answer, but some approximation will be good enough

Veeam server 2008 R2, dual Gigabit connection to LAN, multipath iSCSI on a dedicated network for SAN access
LAN attached Qnap TS-1679U-RP (Veeam accessible dedicated CIFS volume RAID6 made out of 8 SATA disks)

C:\Program Files (x86)\Iperf\Release>Iperf.exe -c 10.0.0.72
------------------------------------------------------------
Client connecting to 10.0.0.72, TCP port 5001
TCP window size: 8.00 KByte (default)
------------------------------------------------------------
[192] local 10.0.0.2 port 58022 connected with 10.0.0.72 port 5001
[ ID] Interval Transfer Bandwidth
[192] 0.0-10.0 sec 1.11 GBytes 949 Mbits/sec

which makes 423,00 GB/h

http://web.forret.com/tools/bandwidth.a ... &unit=Mbps

But my speeds for the backup range from 9 Mb/s to 77 Mb/s ( 32 Gb/h to 277 Gb/h)

Which takes a lot of time when backing up 5.8 Tb server volumes

What one should expect?

Seb

[foggy]

Seb, what transport mode is being effectively used by your jobs and what does bottleneck stats say?

Also please note, that if you are referring to overall job processing rate, then it is just the average speed of processing of all VMs in the job calculated as total amount of processed data divided by the time the job took to finish. So, actually it has nothing to do with link bandwidth at all.

[spgsit5upport]

Ofcourse it has nothing to do with link bandwidth

Let me rephrase the question then...

What speeds (as reported by Veeam in Job Statistics per each disk of the backed up VM) do users achieve for LAN backup to Qnap NAS on the same flat Gigabit network (802.3ad on both server & NAS) using latest Veeam 6.5 & pulling the data from iSCSI dedicated network via multipathing using dedicated iSCSI initiators in Veeam server)

The bottleneck is Source

Makes no difference what iSCSI array is concerned (Dell MD3600i or proper Equallogic PS5000e)
I really find it hard to believe:

14/05/2013 04:57:31 :: Load: Source 98% > Proxy 25% > Network 11% > Target 0%

Seb

[foggy]

Well, have you tried these tips for improving direct SAN mode via iSCSI performance?

[spgsit5upport]

I did, in fact the last post says:

Code: Select all

I changed the Storage Optimized setting for LAN target (in the job properties), and also ran that netsh script at the top of the post....

I just re-ran the job and its now runing at 70MB/s
Equallogic SAN with dedicated iSCSI NIC in the veeam server. (2008 server)

Well, I get more by default, will see if the netsh command (that I completely forgot about) makes difference in this setup...

[dellock6]

Just to add a couple of thoughts:
- link aggregation does not helps too much in a veeam->cifs scenario, both have only 1 IP address and so the IP Hash algorythm only uses 1 channel. Not a real problem anyway until the bottleneck is source and the overall speed does not hit the 125 MB/s limit of a single 1G connection
- what kind of backup mode are you using? I must say the veeam proxy is the physical server itself, right? And what about the job, is it reverse or forward incremental?

Luca.

[dellock6]

Oh, and by the way there is a huge error on the first assumption: IPperf only tests bandwidth, and does not involves disk activity. On a network backup once data cross the wire, they also need to be written to disk (and read from source too)...

IPperf is only saying you that out of a maximum limit of 1 Gbit, you are able to use 949 Mbits, thus giving you 92.67 % efficiency of the network stack.

Luca.