VMs Network speed slowing down

SBG 2
I think u should follow there advice and use the vrack link they mention!

No one here has mentioned this issue on SBG-2, and i bet that if i setup a server on SBG-2 the exact same way - i will have no issues.

We have issues on SBG-1.
 
ok, like to hear the results!
good luck!

ps: u have to place the subnet into the vrack.
 
ok, like to hear the results!
good luck!

ps: u have to place the subnet into the vrack.

An official guide doesn't seem to exist for Proxmox 4.x and vRack is a security issue:
http://pve.proxmox.com/pipermail/pve-devel/2016-January/018793.html

Since it seems you have to do proxy forwarding of internal traffic to the VLAN, something that leaves you vulnerable to customers changing the VM's ip manually to another one in the same IP block - thus taking over other clients traffic.
 
Lets see where they go with this. The puzzling part is one would think this is congestion on the network (at 11 AM ET?) but rebooting the VM fixes it, which throws that idea right out the window. ~_~

Are you still having issues?

I talked to an OVH rep on the phone, they said they would try to have it resolved later today.
 
Are you still having issues?

I talked to an OVH rep on the phone, they said they would try to have it resolved later today.
Nothing new back from OVH as of yet on the ticket after they did a ton of iperf3 tests from various places and escalated to their network techs. So it's a waiting game for me still.
 
Nothing new back from OVH as of yet on the ticket after they did a ton of iperf3 tests from various places and escalated to their network techs. So it's a waiting game for me still.

They didn't do anything about it yesterday even though they said so over the phone and no one touched the ticket.

This is murdering our business...
 
Nothing new back from OVH as of yet on the ticket after they did a ton of iperf3 tests from various places and escalated to their network techs. So it's a waiting game for me still.

They want to do network tests on my production node, would you mind sharing your ticket number so i can save them and us the trouble of doing the same tests again?

EDIT: I sent you a PM
 
Nothing new back from OVH as of yet on the ticket after they did a ton of iperf3 tests from various places and escalated to their network techs. So it's a waiting game for me still.

They said that they cannot test failover IP's, they only test the main host IP - which i've tried to make clear several times now - works just fine.
 
They said that they cannot test failover IP's, they only test the main host IP - which i've tried to make clear several times now - works just fine.
I'll reply/PM you over some extra details in a second. They did happen to update the ticket this morning with showing network congestion "inter-Atlantic" (since my provided test "local host" is the GRA-located EU server)... but I don't know how many times I can state that a reboot on the VM resets those speeds for no apparent reason. Frustrating. I will be testing more this morning from both locations to better reply to them. (NY > BHS and GRA > BHS)
 
I'll reply/PM you over some extra details in a second. They did happen to update the ticket this morning with showing network congestion "inter-Atlantic" (since my provided test "local host" is the GRA-located EU server)... but I don't know how many times I can state that a reboot on the VM resets those speeds for no apparent reason. Frustrating. I will be testing more this morning from both locations to better reply to them. (NY > BHS and GRA > BHS)

Even doing /etc/init.d/networking restart is enough to reset the speeds for 10 minutes.

But yeah, they are simply not listening.
 
Even doing /etc/init.d/networking restart is enough to reset the speeds for 10 minutes.

But yeah, they are simply not listening.
I will need to test that in a KVM-based instance. Right now I'm testing in a container and that command won't restart the network the same way. (hell, I thought containers would be immune to this)
 
This is clearly a router/gateway issue, doing an ARP Cache broadcast towards the OVH Gateway, restores speed again for 7 minutes!

Do this on a VM and see for yourselves:
Code:
arping -c 1 -S VMIP GATEWAYIP

EDIT: arping included in iputils is not the same, i had to do apt-get install arping for it to use the right one.
 
Last edited:
This is clearly a router/gateway issue, doing an ARP Cache broadcast towards the OVH Gateway, restores speed again for 7 minutes!

Do this on a VM and see for yourselves:
Code:
arping -c 1 -S VMIP GATEWAYIP

EDIT: arping included in iputils is not the same, i had to do apt-get install arping for it to use the right one.

Code:
-----------------------------------------------------------
Server listening on 5201
-----------------------------------------------------------
Accepted connection from , port 25425
[  5] local  port 5201 connected to  port 53203
[ ID] Interval           Transfer     Bandwidth
[  5]   0.00-1.00   sec  1.78 MBytes  14.9 Mbits/sec
[  5]   1.00-2.00   sec  1.23 MBytes  10.3 Mbits/sec
[  5]   2.00-3.00   sec  1.20 MBytes  10.1 Mbits/sec
[  5]   3.00-4.00   sec  1.59 MBytes  13.3 Mbits/sec
[  5]   4.00-5.00   sec  1.33 MBytes  11.2 Mbits/sec
[  5]   5.00-6.00   sec   785 KBytes  6.43 Mbits/sec
[  5]   6.00-7.00   sec   827 KBytes  6.78 Mbits/sec
[  5]   7.00-8.00   sec   953 KBytes  7.81 Mbits/sec
[  5]   8.00-9.00   sec  1.40 MBytes  11.8 Mbits/sec
[  5]   9.00-10.00  sec  1.55 MBytes  13.0 Mbits/sec
[  5]  10.00-11.00  sec  1.46 MBytes  12.2 Mbits/sec
[  5]  11.00-12.00  sec   856 KBytes  7.01 Mbits/sec
[  5]  12.00-13.00  sec   727 KBytes  5.95 Mbits/sec
[  5]  13.00-14.00  sec   658 KBytes  5.39 Mbits/sec
[  5]  14.00-15.00  sec  1.46 MBytes  12.2 Mbits/sec
[  5]  15.00-16.00  sec  2.37 MBytes  19.9 Mbits/sec
-snip-
arping -c 1 -s vm_ip gw_ip
-/snip-
[  5]  16.00-17.00  sec  2.66 MBytes  22.3 Mbits/sec
[  5]  17.00-18.00  sec  2.65 MBytes  22.2 Mbits/sec
[  5]  18.00-19.00  sec  2.63 MBytes  22.1 Mbits/sec
[  5]  19.00-20.00  sec  2.62 MBytes  22.0 Mbits/sec
[  5]  20.00-21.00  sec  2.66 MBytes  22.3 Mbits/sec
[  5]  21.00-22.00  sec  2.70 MBytes  22.7 Mbits/sec
[  5]  22.00-23.00  sec  2.51 MBytes  21.0 Mbits/sec
[  5]  23.00-24.00  sec  2.68 MBytes  22.5 Mbits/sec
[  5]  24.00-25.00  sec  2.58 MBytes  21.6 Mbits/sec
[  5]  25.00-26.00  sec  2.63 MBytes  22.1 Mbits/sec
[  5]  26.00-27.00  sec  2.64 MBytes  22.2 Mbits/sec
[  5]  27.00-28.00  sec  2.57 MBytes  21.6 Mbits/sec
[  5]  28.00-29.00  sec  2.63 MBytes  22.0 Mbits/sec
[  5]  28.00-29.00  sec  2.63 MBytes  22.0 Mbits/sec

Looks like that did it for me. I'll be updating the ticket with OVH on this to see what they say.
 
Last edited:
I just want to point out that this is only a temporary fix that OVH have applied, if you change the server that the IP range is bound to - the speed will again be slow after 7 minutes.
 
Just joining in on the issue. I was in the process of migrating 40 nodes to a new box with them from our UK supplier, luckily I noticed this issue with the first VM I put on there. I will also be creating a ticket.
 
Later 3 months with the issue i left the server. I gave it other opportunity to ovh and i rent another server in another datacenter.

I buy another OVH server (in Roubaix) in this case, and i move my current IPs (a /27 and a /28 subnet). And yes, working without issues, absolutely, at this time! And of course with the same IPs And i hope that this continues well.

It is clearly that is their issue, due "certain" routers in "certain" locations/routes are bad and/or overloaded and do nothing, simply.

Over this past 3 moths talking with OVH regarding issue, and sending all kind of logs and evidences, the always remit me reboot in rescue mode and test on my main IP, of course i can´ t set vmac for additional IPs on rescue and of course they don´t take care of this, since there are a virtualization "software" and they remit me again to rescue mode and test.

So it appears stupid, but i could think that they knows perfectly the issue, but they really not want upgrade their routers and/or infraestructure in this case, it is expensive.

I can not understand in any way case omise all of clients giving this results and evidences. And sure maybe hundred or thousands ovh customers with the issue, but they don´t make test, simply "works" and is enough for them.

I´m sorry for this words but i have to say.

Now, all apart of this, in the new server (more expensive) in Roubaix working like a charm since 4 days.

So resuming: another server and good luck!


Greetings!
 

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!