Both 1Gbit and 10 Gbit, avoid throttle ?

[lars@norstat.no]

Hi, everyone. Haven't been on the forums for a while since everything has been running stable for a while now. However i have upgraded my system somewhat and have run into a problem.

I now have 2x8 Gbit FC between my Datacenter and DR site and 2x10 Gbit between my Veeam Proxy and DR ESXi host. Still i'm only getting around 500 mbit troughput.

On the target site i have bought a new Storwize v5000 to use for destination for replica and backup is stored localy on the Veeam proxy.

when testing i get around 14 -16 Gbit between the sites so the FC is working great. Storage vmotion is fast and everything.

Now i think i know what the throttle is about, because i have read several places that NBD traffic is limited when running on 1 Gbit nic, but should not be limited when running on a 10 Gbit nic.

Both my Veeam Proxy and DR ESXi host are running a mix of 1 Gbit and 10 Gbit adapters. Could there be a problem with Vmware throttling because i have a 1 Gbit adapter in the machine ?

The 1 Gbit network does have a Vmkernel port that is used for management. And there are two seperate vmkernel ports connected to the 10 Gbit adapters.

Even though i have entered the correct ip addresses for the ESXi host in the HOST file of the Veeam Proxy so and that i have confirmed that the data is indeed flowing over the 10 Gbit network i still get the message "no available proxies are running on ESXi management interface subnet".

But as i say the traffic is taking the correct route.

How do i disable this cap ?

[Vitaliy S.]

Hello Lars,

Both my Veeam Proxy and DR ESXi host are running a mix of 1 Gbit and 10 Gbit adapters. Could there be a problem with Vmware throttling because i have a 1 Gbit adapter in the machine ?

I might have missed it, but is it a virtual or physical proxy server?

Even though i have entered the correct ip addresses for the ESXi host in the HOST file of the Veeam Proxy so and that i have confirmed that the data is indeed flowing over the 10 Gbit network i still get the message "no available proxies are running on ESXi management interface subnet".

What about Veeam backup server? I believe this verification is performed by the server, and not by the proxy.

Thank you!

[lars@norstat.no]

Thnx for your answer Vitaliy and i'm sorry for taking such long time to answer.

I have now found out that the bottleneck must be on the fiber end of the Veeam Server and i'm gonna make a new post.

[Vitaliy S.]

Ok, but it still not clear about proxy resources availability. Did you make changes to the hosts file on the backup server?

[lars@norstat.no]

Yea, about that. I can't change the host file on the Veeam Backup server because that server can't contact the ESXi host on that network.

Veeam -> ESXi Target via 1 Gbit management network
Veeam -> Veeam Proxy via 1 Gbit management network
Veeam Proxy -> ESXi Target via CLOSED 10 Gbit network

Changing the host file on the Veeam Proxy ensures that the data is flowing in the right direction, but the message still appears. More of an irritation or level 999 bug

[Andreas Neufert]

ALL veeam Servers need to be able to lookup all EsXi hosts and vcenter onshor and full qualified domain names.
Regardless if iP s are used to add something at veeam, because vmware api uses names fqdn in the background with layest vmware versions.


You can use host file modification on all hosts.
Take care that the ip addresses you enter are bound to a vmkernel interface at host that can process our backup traffic. there are some options at vmware to prevent that.

At veeam proxy configuration you need to the network mode or at least fall back to network mode. Without this a replication is not possible if hotadd proxy on target is not available.

Ping the names short and long of all esxi hosts and check if all hosts answer.

[lars@norstat.no]

Actually this works without the Veeam management server not beeing able to see all the interfaces on the Target ESXi host, but still the message appears but it's no big deal

i can't add the 10 Gbit mangement kernel to the host file of the Veeam Management server because that server can't reach that interface.

[Andreas Neufert]

Then this onfiguration is not supported, because we send directly to ESXi host commands from the B&R Server itself. At least you need a routet access to the interface on port 443 and is some cases on 902.