Comprehensive data protection for all workloads
Post Reply
hanierafatmanesh
Novice
Posts: 4
Liked: never
Joined: Jan 18, 2020 8:12 pm
Full Name: Hanieh Rafatmanesh
Contact:

Low Speed on 10Gb Network

Post by hanierafatmanesh » Jan 18, 2020 8:28 pm

Hi,
We are two servers in our company that is connected each other via 10Gb NIC and cable for seprating backup network with high speed.test is done for vmware backup job and backup from another Physical host with agent.but maximum speed is  1-2Gb.I dont know where is bottleneck.It is worth mentioning:
1-File transfer between two servers is up to 7Gb so that resource performance such as hard speed and CPU core is ok!
2-we checked throttling and network rule that not to set.
Sincerely,
Hanieh Rafatmanesh

foggy
Veeam Software
Posts: 18818
Liked: 1655 times
Joined: Jul 11, 2011 10:22 am
Full Name: Alexander Fogelson
Contact:

Re: Low Speed on 10Gb Network

Post by foggy » Jan 19, 2020 10:06 pm

Hi Hanieh, what is reported as the bottleneck by the backup job itself? You can see it in the job session window.

Virtuollie
Service Provider
Posts: 13
Liked: 6 times
Joined: Jan 03, 2020 10:08 am
Full Name: Oliver Palz
Contact:

Re: Low Speed on 10Gb Network

Post by Virtuollie » Jan 19, 2020 10:29 pm

Did you try allowing more threads on the backup proxy?
How many did you configure?

hanierafatmanesh
Novice
Posts: 4
Liked: never
Joined: Jan 18, 2020 8:12 pm
Full Name: Hanieh Rafatmanesh
Contact:

Re: Low Speed on 10Gb Network

Post by hanierafatmanesh » Mar 22, 2020 5:40 pm

Hi
Sorry for late replying
My backup is in one VM on ESXi with 10Gb network and in the job session window there is 99% Source and 10% Network & Proxy.Although my VM source is suitable.
Unfortunately My problem persists and maximum speed is 2Gbps.

hanierafatmanesh
Novice
Posts: 4
Liked: never
Joined: Jan 18, 2020 8:12 pm
Full Name: Hanieh Rafatmanesh
Contact:

Re: Low Speed on 10Gb Network

Post by hanierafatmanesh » Mar 22, 2020 5:48 pm

Hi
Sorry for late replying
I don't separate Proxy server, but I checked factors such as:
Transport mode is on: Network
Concurrent job is on: 1
I/O control is disable

PetrM
Veeam Software
Posts: 362
Liked: 49 times
Joined: Aug 28, 2013 8:23 am
Full Name: Petr Makarov
Location: Prague, Czech Republic
Contact:

Re: Low Speed on 10Gb Network

Post by PetrM » Mar 22, 2020 9:50 pm

Hi Hanieh,

"99 % Source" means that data retrieving in the "Network" mode is the slowest data processing stage.
The speed in "Network" mode can be decreased because of the two main reasons:
1. High read latency on the source disk sub-system
2. Bad performance of network channel between ESXi and proxy (likely not the reason in our case)

I would try to test Hot Add mode and to specify preferred network setting to guarantee that data traffic goes from the virtual proxy to the repository over the fast backup network.
And one more consideration is to increase number of concurrent tasks according to available proxy resources.

I would inspect disk read performance if the result of test was the same or worse, for example ESXTOP tool might be useful.

Thanks!

hanierafatmanesh
Novice
Posts: 4
Liked: never
Joined: Jan 18, 2020 8:12 pm
Full Name: Hanieh Rafatmanesh
Contact:

Re: Low Speed on 10Gb Network

Post by hanierafatmanesh » Mar 26, 2020 10:16 pm

Hi PetrM
1.Read speed on source VM is 7-8Gbps that is test with Crystaldiskmark tools(Hard R&W speed tools) and Hard Write speed on Veeam server is 9Gbps
2.I tested this one(Hot add mode) and dedicate preferred network but it was no different.
3.I changed concurrent task from one to two and speed grew up a little from 1.5 to 2.1Gbps. but other increase of concurrent number was useless.
So i think this problem is only veeam software. and my question is that without any change on veeam software it could be take backup with 10Gbps?
Thanks for your response

PetrM
Veeam Software
Posts: 362
Liked: 49 times
Joined: Aug 28, 2013 8:23 am
Full Name: Petr Makarov
Location: Prague, Czech Republic
Contact:

Re: Low Speed on 10Gb Network

Post by PetrM » Mar 27, 2020 12:11 am

Hi Hanieh,

I think we should not compare results of the mentioned benchmark tools running directly on VM with processing rate reported by Veeam job because mechanisms of data read are totally different in both cases.

If you would like to perform benchmark tests in right way, I strongly recommend to contact our support team.
They have special tools to emulate data read in similar way as Veeam does during backup (depending on transport mode). Please do not forget to paste support case id for our reference.

Thanks!

Post Reply

Who is online

Users browsing this forum: Hiroyuki Uemura and 31 guests