OVH Community, your new community space.

Horrible Network speed on 2x new servers in 94.23.49 range


Criot
05-01-2015, 13:13
Quote Originally Posted by mgould73
If it its?? and where did you hear about this? and its only effecting certain servers?

I mean, I still don't get it. Unless these 94.23.49.* servers are running off a completely different switch and routed differently that is the only explanation I can come up with.

I have no testing prior on these, because they are new servers. Only delivered on the 31st, but have been this way since first testing.
Oles mentioned it on Twitter, there's several statuses about VAC issues on status.ovh.net as well.

mgould73
05-01-2015, 02:37
Quote Originally Posted by Criot
From what I gather OVH have been under a heavy DDoS over the last few days, the most likely cause would be that.
If it its?? and where did you hear about this? and its only effecting certain servers?

I mean, I still don't get it. Unless these 94.23.49.* servers are running off a completely different switch and routed differently that is the only explanation I can come up with.

I have no testing prior on these, because they are new servers. Only delivered on the 31st, but have been this way since first testing.

Criot
05-01-2015, 02:18
Quote Originally Posted by mgould73
Just a follow up to this, it seems the entire OVH Network has gone to hell the past few days. No Answers to any of my tickets, and the speed b/w problems continue.

Can someone care to explain why a server on the 94.23.214.* range is 10x faster then the 94.23.49.* range a rack away in the same data center?

I've posted my traceroutes above in first post, but I just don't get it.

Using the same EXACT software on all 3 servers, server in range 94.23.214.* completely maxes out my home 100mbit connection over rsync, VSFTP or any other protocol. Yet my 2 new servers in 94.23.49.* range I cannot break 10mbit??? The server are in the same RBX2 data center and a rack away.

Even OVH-OVH is crap. Here is the BEST speedtest I can run (upstream) from one of the 94.23.49 servers.

Hosted by FreeMobile (Paris) [1.59 km]: 5.063 ms
Testing download speed........................................
Download: 426.71 Mbits/s
Testing upload speed............................................. .....
Upload: 77.40 Mbits/s

Here is my server on 94.23.214.* range

Selecting best server based on latency...
Hosted by FreeMobile (Paris) [1.59 km]: 4.968 ms
Testing download speed........................................
Download: 617.80 Mbits/s
Testing upload speed............................................. .....
Upload: 247.74 Mbits/s

Anyone care to chime in on why there is such a huge difference? I would like support to look at this as well, because I don't understand how my 2x new server seem to run like adsl speed and my other is perfectly fine.
From what I gather OVH have been under a heavy DDoS over the last few days, the most likely cause would be that.

mgould73
05-01-2015, 01:53
Just a follow up to this, it seems the entire OVH Network has gone to hell the past few days. No Answers to any of my tickets, and the speed b/w problems continue.

Can someone care to explain why a server on the 94.23.214.* range is 10x faster then the 94.23.49.* range a rack away in the same data center?

I've posted my traceroutes above in first post, but I just don't get it.

Using the same EXACT software on all 3 servers, server in range 94.23.214.* completely maxes out my home 100mbit connection over rsync, VSFTP or any other protocol. Yet my 2 new servers in 94.23.49.* range I cannot break 10mbit??? The server are in the same RBX2 data center and a rack away.

Even OVH-OVH is crap. Here is the BEST speedtest I can run (upstream) from one of the 94.23.49 servers.

Hosted by FreeMobile (Paris) [1.59 km]: 5.063 ms
Testing download speed........................................
Download: 426.71 Mbits/s
Testing upload speed............................................. .....
Upload: 77.40 Mbits/s

Here is my server on 94.23.214.* range

Selecting best server based on latency...
Hosted by FreeMobile (Paris) [1.59 km]: 4.968 ms
Testing download speed........................................
Download: 617.80 Mbits/s
Testing upload speed............................................. .....
Upload: 247.74 Mbits/s

Anyone care to chime in on why there is such a huge difference? I would like support to look at this as well, because I don't understand how my 2x new server seem to run like adsl speed and my other is perfectly fine.

heise
31-12-2014, 00:40
Why should I need to reroute traffic
I am sure that solution is faster and less painful.
I recommend you read http://forum.ovh.co.uk/showthread.ph...s-an-emergency

Welcome to the wonderful SYS world....

mgould73
30-12-2014, 18:19
Quote Originally Posted by heise
This has no impact what so ever on the internet speed. ICMP are low priority and when the server gets busy, they just drop them. Nothing wrong with that. See unoffical FAQ on detecting network problems.

Why don't you route traffic manually over the SYS server (use like proxy) that is fast?

Why should I need to reroute traffic using a proxy or anything else for that matter when there is clearly a problem?
My other server at the same DC on different rack/switch has no problems, my server at SBG and GRA have no problems, but these 2 do.

and the common denominator I can see being in the same DC is that main switch these 2x new servers are on.

heise
30-12-2014, 18:11
10 2210 ms * * rbx-g1-a9.fr.eu [91.121.131.189]
This has no impact what so ever on the internet speed. ICMP are low priority and when the server gets busy, they just drop them. Nothing wrong with that. See unoffical FAQ on detecting network problems.

Why don't you route traffic manually over the SYS server (use like proxy) that is fast?

mgould73
30-12-2014, 17:46
Looking at the weathermap, vss-1-6k is in the dark orange which is where my 2 problem servers are it seems. I wonder if they have any plans to reroute some traffic differently. vss-2-6k is green. (my other server)

mgould73
30-12-2014, 17:42
Quote Originally Posted by Smally1997
It does appear that anything going through rbx-g1-a9 is experiencing high packet loss right now.
Possibly, but my other sys server that goes through rbx-g1-a9 as well still continues to perform well.

I don't know if this is a new problem in this range or what, as I only received these 2x new servers 2 days ago. But since they have been online, both have been like this for 2 days.

Averaging about 5-15mbit on the upload, where as my other server to the same exact places will reach its maximum 250mbit all the time.
So what's the difference? Servers are in the same DC RBX2, just on a different rack/switch it seems.

Smally1997
30-12-2014, 17:19
It does appear that anything going through rbx-g1-a9 is experiencing high packet loss right now.

mgould73
30-12-2014, 17:03
Hi, since no answer from support yet I figure I would see if anyone else has any issues here.

I have 1 SYS server ESAT1 in the 94.23.214.* range. The network performance to .eu .us is great, traceroute shows this range of IP is on switch vss-2-6k. Max speed on DL/UP.

I Ordered 2x more SYS servers ESAT2, I was given the range 94.23.49.* This range is awful, and only difference I can tell is that this is on router vss-1-6k. Performance only seems ok to my other OVH servers and sometimes that's hit or miss with data speeds dropping less then 50mbit. I don't have this problem on the 94.23.214.* range. (problem seems to be the upstream only)

Testing backups with a 10gb database file I can get max upstream on 94.23.214 range, but the 94.23.49 range transfers 1-10mbit to the same location. OVH-Internet (both new servers) I also seem to see huge ping on rbx-g1-a9 over 2000ms.


Tracert to 1 of the new SYS servers.

1 <1 ms <1 ms <1 ms 192.168.1.1
2 * * * Request timed out.
3 12 ms 9 ms 12 ms 433be075.cst.lightpath.net [67.59.224.117]
4 34 ms 12 ms 11 ms ool-4353fb8d.dyn.optonline.net [67.83.251.141]
5 16 ms 11 ms 11 ms 65.19.119.209
6 11 ms 14 ms 11 ms r1-gig1-0-0.in.nycmnyzr.cv.net [64.15.0.30]
7 * * * Request timed out.
8 * * * Request timed out.
9 * * * Request timed out.
10 2210 ms * * rbx-g1-a9.fr.eu [91.121.131.189]
11 107 ms 108 ms 92 ms vss-1-6k.fr.eu [94.23.122.169]
12 107 ms 90 ms 92 ms ns301315.ip-94-23-49.eu

Tracert to my 94.23.214 Server that does not have any problems (note the still very high ping on rbx-g1-a9)
1 <1 ms <1 ms <1 ms 192.168.1.1
2 * * * Request timed out.
3 9 ms 22 ms 9 ms 433be075.cst.lightpath.net [67.59.224.117]
4 15 ms 11 ms 9 ms ool-4353fb8d.dyn.optonline.net [67.83.251.141]
5 11 ms 11 ms 11 ms 64.15.7.49
6 11 ms 12 ms 12 ms 64.15.3.250
7 * * * Request timed out.
8 40 ms * 32 ms nwk-1-6k.nj.us [178.32.135.57]
9 * * * Request timed out.
10 2254 ms * 2235 ms rbx-g1-a9.fr.eu [178.33.100.65]
11 92 ms 205 ms * vss-2-6k.fr.eu [94.23.122.173]
12 93 ms 91 ms 95 ms ns304751.ip-94-23-214.eu

Only real Difference I can see is the vss-1-6k in the new servers and vss-2-6k on my older SYS server which has 0 problems. Anyone else within that range that is having very bad upstream problems?

At the moment they are both completely useless as it is now taking hours instead of a few minutes to transfer data. Only thing I can do is send the backup databases to the good working range and use that server to upload, but that is a complete waste of $70 .eu. I would have thought the performance would be the same.