OVH Community, your new community space.

RBX4: vss-5a-6k and vss-5b-6k


Loic
23-03-2012, 10:00
The high latency variation was in my case not the fact of OVH, my mistake, but from Google, the fact that the issues arrived at the same time sent me the wrong way. Stupidly, my instance running the HTTP probes had no local DNS cache and was hitting the Google DNS for each request...

Here are the changes, first well, then Google going bad and me with them, then using a local to the instance DNS cache:



I wrote a bit more about this issue with a way to check if you are affected. I really felt stupid

Loic
13-03-2012, 07:56
You can see the work details, they are active this morning.

William_GL
09-03-2012, 16:24
If it makes any difference the servers we're having issues are on sw-n5-118.248/249

I've just noticed that both these switches and sw-n5-226.248/249 have had additional uplinks added in the last 24 hours or so, but they still appear as inactive at present on the weathermap - is this an indication of progress?

William_GL
09-03-2012, 16:10
Still getting major issues here.

Connectivity problems between two servers on the same switch under these routers. This is happening on a daily basis, but worse today - lost connectivity for almost 30 seconds twice today.

I stupidly thought that getting all our servers with the same company, in the same data centre would guarantee no connection issues between them.

When will this be fixed OVH?

Do you think this level or service is acceptable?

Not trying to be cheeky here, would rather you were honest if you've no intention of addressing this issue for whatever reason, as these are mission critical servers and we will move them elsewhere if this is the case.

Loic
06-03-2012, 14:23
It looks like this is a new customer or a change in behaviour of a current customer. Here: http://weathermap.ovh.net/roubaix-4 on the switch sw-n5-226.249 exactly like yesterday, above 70% usage. Not as bad as yesterday, but starting around 9 UTC, lot of erratic and variance in the request time. I have they will find who is doing that and adjust, because if one user is destroying the experience of others, this is not good.

William_GL
05-03-2012, 17:50
Yup, we're in the same boat where we've applications on the different servers relying on good connectivity between each other (databases, web servers, voip systems), and it's annoying when connectivity on the internal OVH network isn't 100%.

I've had to start adjusting receive/transmit buffers and other things, which were initially tweaked for servers on a local network to each other, but unfortunately we've still some applications which are just very intolerant of packet loss.

Loic
05-03-2012, 16:51
Since 15h10 UTC it is improving, but still a lot of deviation for some points. In the big scheme at the "internet" level, it is not "that" bad, but for internal traffic OVH to OVH, this is annoying as I rely on a good network between my severs (DB on one and app on another).

Loic
05-03-2012, 15:32
Today is really really bad, the 3 lines at the bottom are internal to OVH, the blue is with Gandi (Paris). This is the response time of a HTTP GET request on the home page.



Easy to spot that this is not a normal situation and I do hope they are going to fix it asap.

William_GL
05-03-2012, 13:08
I've 4 servers connected to these and have been experiencing issues for weeks now, including low speeds, drops in connectivity and packet loss.

Even getting packet loss between the servers, so it's definitely these to blame.

http://status.ovh.net/?do=details&id=2337

Seems like OVH are aware of the issue but it's either not a high priority or they're just not keeping us updated of what progress is being made.

mike_
04-03-2012, 17:45
These seem to be under extremely high load a lot recently. I'm only getting 1Mbit/sec from my server. Any plans to fix?