We are in the process of migrating this forum. A new space will be available soon. We are sorry for the inconvenience.

Buffering issues


Andy
19-03-2008, 11:41
Quote Originally Posted by milanyc
Andy, have you in reality EVER achieved 100mbit upload speed from a single SLA ovh server?
Note that I am not talking theory here but the actual experienced performance.
Yes, lots of times. See the post I made in the sla/nosla thread.

milanyc
19-03-2008, 11:20
Andy, have you in reality EVER achieved 100mbit upload speed from a single SLA ovh server?
Note that I am not talking theory here but the actual experienced performance.

Andy
19-03-2008, 11:17
Quote Originally Posted by milanyc
and which restrictions apply to a user with only 1 SLA server?
Dont tell me they get 100mbit ALL THE TIME...

If your server is on SLA you are entitled to 100Mbps 24/7, no limits.


Quote Originally Posted by kurtis08
all myne are sla and i cant hit 20meg

thanks
How are you testing this? Give as much detail as possible for the best chance of us helping you.

What is the speed of the connection you're transferring from/to?
Where is it based?
How many connections did you use?

kurtis08
19-03-2008, 04:55
all myne are sla and i cant hit 20meg

thanks

milanyc
18-03-2008, 23:53
and which restrictions apply to a user with only 1 SLA server?
Dont tell me they get 100mbit ALL THE TIME...

BELLonline
18-03-2008, 22:57
I've been looking for info on the bandwidth policy and it looks like once your total usage goes above 101Mbit/s your servers are put in to noSLA mode:

To sum up , noSLA is:
- Total bandwidth per server is not limited. You are entitled to 100Mbps or 1Gbps (according to the offer).
- Bandwidth per connection
- Between 0h and 9am: 10 Mbps per connection
- Between 9am and 3pm: reduce of 5 Mbps to 500 Kbps per connection
- Between 3 pm and 9pm: 500 Kbps per connection
- Between 9pm and 11pm: re-increase 500Mbps to 5Mbps per connection
- Between 11pm and midnight: 10 Mbps per connection

kurtis08
18-03-2008, 22:41
i found out something pretty special today

that ovh apparently limit your while account to 101mbps

so if i have 10 servers

i have to make sure for them not to buffer they all equal 101mbps in total

so say if one server uses 50mbps

then i have 50mbps to share between the other 9



and im sure its not a config error as i run this config on 9 other servers in the world in uk and usa and fine

thanks
kurtis

BELLonline
18-03-2008, 21:50
I recommend this, I use it all the time if I believe that one of my servers is experiencing a problem http://www.pingplotter.com/

You can see exactly what carrier is causing problems or if it is your server.

Andy
18-03-2008, 16:54
Quote Originally Posted by kurtis08
but what im saying is

why have a bandwidth carrier like global crossing

if you cant even make a reliable connection with the usa

and its not just usa buffering now

i got poland buffering spain,germany and a few others

so its your network
Ever thought it could be your software or server? I haven't got any transit issues to anybody that visit my sites or download files from me.

kurtis08
18-03-2008, 16:17
but what im saying is

why have a bandwidth carrier like global crossing

if you cant even make a reliable connection with the usa

and its not just usa buffering now

i got poland buffering spain,germany and a few others

so its your network

Andy
18-03-2008, 09:59
Quote Originally Posted by oles@ovh.net
> but yet you use globalcrossing who have one of the biggest networks in
> america


gblx, teleglobe, level3 are the great networks, but it costs lot of
money. in place to pay this money, I prefer to invest the money in
my own network in europe and try to peer with the customers of gblx,
teleglobe and level3. in 2-3 years it will cost less money.

selling the server in the country that we have no this kind of strategy
it means that we will pay lot of money for the bandwidth with no hope
to earn some. for example 100Mbps costs 1000Euro/mo. kimsufi is sold
20euro/mo with 100Mbps. it's possible because we have our own network.

Wow, bandwidth costs a lot... At least I now know why so few companies can offer 100Mbps unmetered bandwidth...

oles@ovh.net
18-03-2008, 07:14
> but yet you use globalcrossing who have one of the biggest networks in
> america


gblx, teleglobe, level3 are the great networks, but it costs lot of
money. in place to pay this money, I prefer to invest the money in
my own network in europe and try to peer with the customers of gblx,
teleglobe and level3. in 2-3 years it will cost less money.

selling the server in the country that we have no this kind of strategy
it means that we will pay lot of money for the bandwidth with no hope
to earn some. for example 100Mbps costs 1000Euro/mo. kimsufi is sold
20euro/mo with 100Mbps. it's possible because we have our own network.


kurtis08
17-03-2008, 19:07
Quote Originally Posted by Marco
The OVH network is peered with level3 in Europe. That does not mean that the rest of the world is cut off. That means that OVH's network in Europe is shared among other peered companies, those which are not peered, will not have a direct route to us. OVH uses Level3's Tier 1 network (Level 3 operates one of the largest backbone networks of any internet service provider), therefore OVH grants great accessibility from all over the world, especially in Europe. The rest of the world has normal access to our network, that means that if in US some people do not have a well routed connection, they might have a slower connection to us. That does not happen if they get a normal connection to the internet.
but yet you use globalcrossing who have one of the biggest networks in america

something doesnt make sense

Marco
17-03-2008, 18:56
The OVH network is peered with level3 in Europe. That does not mean that the rest of the world is cut off. That means that OVH's network in Europe is shared among other peered companies, those which are not peered, will not have a direct route to us. OVH uses Level3's Tier 1 network (Level 3 operates one of the largest backbone networks of any internet service provider), therefore OVH grants great accessibility from all over the world, especially in Europe. The rest of the world has normal access to our network, that means that if in US some people do not have a well routed connection, they might have a slower connection to us. That does not happen if they get a normal connection to the internet.

kurtis08
17-03-2008, 16:56
so basically ovh is cut off to countries which they dont list,

and yet they have 160gbps network and if one of the biggest countries cant get to the network whats the point of having a big network

anyway ill cut my loses and move to a premium network more money but worth it at least my servers are available world wide

Allot
14-03-2008, 18:34
Quote Originally Posted by Andy
So that statement basically says that you can't provide hosting in one place for everyone in the world because your network won't let us!

You have huge pipes with peering to 10's maybe hundreds of networks, and you can't guarantee you can connect to them. Yet I have residential broadband for a fraction of the cost, fraction of the pipe size, and I have no issues connecting to anywhere in the world.

How very incompetent. I've never heard such rubbish from a data center.

On every dedicated server product page, you can find the following information where it says "Available" :

http://www.ovh.co.uk/items/cadre/disponibilite.xml

These are the countries we provide to and give an SLA guarantee. We are working on getting more connections to a greater number of countries (see threads like "link : Roubaix/Bruxelles/Amsterdam"). Our aim is to give a high quality connection to the countries we provide to and to maintain this quality as our network grows.


Allot

Andy
14-03-2008, 18:00
Quote Originally Posted by Allot
See the reason why OVH offers servers to a certain number of countries is because the connection can be guaranteed to them. The rest of the countries cannot be guaranteed (hence the problems you are facing) and that's why its not supplied there.

Allot
So that statement basically says that you can't provide hosting in one place for everyone in the world because your network won't let us!

You have huge pipes with peering to 10's maybe hundreds of networks, and you can't guarantee you can connect to them. Yet I have residential broadband for a fraction of the cost, fraction of the pipe size, and I have no issues connecting to anywhere in the world.

How very incompetent. I've never heard such rubbish from a data center.

Allot
14-03-2008, 17:46
Quote Originally Posted by kurtis08
it isnt a server issue its basically there network

there network isnt working good internationally thats why they dont want to respond
See the reason why OVH offers servers to a certain number of countries is because the connection can be guaranteed to them. The rest of the countries cannot be guaranteed (hence the problems you are facing) and that's why its not supplied there.

Allot

kurtis08
14-03-2008, 17:37
it isnt a server issue its basically there network

there network isnt working good internationally thats why they dont want to respond

Allot
14-03-2008, 17:19
Quote Originally Posted by Andy
I am able to ping 2 of your 3 servers.

I can access your servers just fine (well, on the one that responds to HTTP I can).

I'd like to help you as much as possible with this problem since OVH don't seem to want to.
I have received normal ping results on all three of your servers. To clarify, do you only have problems connecting from the US?

Allot

stationX:~# ping ns28159.ovh.net
PING ns28159.ovh.net (91.121.91.149) 56(84) bytes of data.
64 bytes from s3.rentaserv.net (91.121.91.149): icmp_seq=1 ttl=47 time=64.3 ms
64 bytes from s3.rentaserv.net (91.121.91.149): icmp_seq=2 ttl=47 time=62.8 ms
64 bytes from s3.rentaserv.net (91.121.91.149): icmp_seq=3 ttl=47 time=65.9 ms
64 bytes from s3.rentaserv.net (91.121.91.149): icmp_seq=4 ttl=47 time=58.9 ms
64 bytes from s3.rentaserv.net (91.121.91.149): icmp_seq=5 ttl=47 time=66.2 ms
64 bytes from s3.rentaserv.net (91.121.91.149): icmp_seq=6 ttl=47 time=72.8 ms
64 bytes from s3.rentaserv.net (91.121.91.149): icmp_seq=7 ttl=47 time=72.3 ms
64 bytes from s3.rentaserv.net (91.121.91.149): icmp_seq=8 ttl=47 time=92.5 ms
64 bytes from s3.rentaserv.net (91.121.91.149): icmp_seq=9 ttl=47 time=64.2 ms
64 bytes from s3.rentaserv.net (91.121.91.149): icmp_seq=10 ttl=47 time=63.3 ms

[11]+ Stopped ping ns28159.ovh.net
stationX:~# ping ns28189.ovh.net
PING ns28189.ovh.net (91.121.92.19) 56(84) bytes of data.
64 bytes from s6.rentaserv.net (91.121.92.19): icmp_seq=1 ttl=44 time=78.3 ms
64 bytes from s6.rentaserv.net (91.121.92.19): icmp_seq=2 ttl=44 time=64.1 ms
64 bytes from s6.rentaserv.net (91.121.92.19): icmp_seq=3 ttl=44 time=59.9 ms
64 bytes from s6.rentaserv.net (91.121.92.19): icmp_seq=4 ttl=44 time=64.3 ms
64 bytes from s6.rentaserv.net (91.121.92.19): icmp_seq=6 ttl=44 time=64.3 ms
64 bytes from s6.rentaserv.net (91.121.92.19): icmp_seq=9 ttl=44 time=61.9 ms
64 bytes from s6.rentaserv.net (91.121.92.19): icmp_seq=10 ttl=44 time=68.7 ms
64 bytes from s6.rentaserv.net (91.121.92.19): icmp_seq=11 ttl=44 time=59.9 ms
64 bytes from s6.rentaserv.net (91.121.92.19): icmp_seq=13 ttl=44 time=61.6 ms
64 bytes from s6.rentaserv.net (91.121.92.19): icmp_seq=14 ttl=44 time=60.7 ms
64 bytes from s6.rentaserv.net (91.121.92.19): icmp_seq=15 ttl=44 time=63.6 ms

[12]+ Stopped ping ns28189.ovh.net
stationX:~# ping ns27361.ovh.net
PING ns27361.ovh.net (91.121.85.17) 56(84) bytes of data.
64 bytes from s7.rentaserv.net (91.121.85.17): icmp_seq=1 ttl=47 time=60.3 ms
64 bytes from s7.rentaserv.net (91.121.85.17): icmp_seq=2 ttl=47 time=65.2 ms
64 bytes from s7.rentaserv.net (91.121.85.17): icmp_seq=3 ttl=47 time=62.4 ms
64 bytes from s7.rentaserv.net (91.121.85.17): icmp_seq=4 ttl=47 time=61.5 ms
64 bytes from s7.rentaserv.net (91.121.85.17): icmp_seq=5 ttl=47 time=62.6 ms
64 bytes from s7.rentaserv.net (91.121.85.17): icmp_seq=6 ttl=47 time=59.7 ms
64 bytes from s7.rentaserv.net (91.121.85.17): icmp_seq=8 ttl=47 time=58.1 ms
64 bytes from s7.rentaserv.net (91.121.85.17): icmp_seq=9 ttl=47 time=58.9 ms
64 bytes from s7.rentaserv.net (91.121.85.17): icmp_seq=10 ttl=47 time=60.1 ms
64 bytes from s7.rentaserv.net (91.121.85.17): icmp_seq=11 ttl=47 time=59.2 ms

[13]+ Stopped ping ns27361.ovh.net

Andy
14-03-2008, 14:31
I am able to ping 2 of your 3 servers.

C:\Documents and Settings\Andrew>ping ns28159.ovh.net

Pinging ns28159.ovh.net [91.121.91.149] with 32 bytes of data:

Reply from 91.121.91.149: bytes=32 time=57ms TTL=48
Reply from 91.121.91.149: bytes=32 time=54ms TTL=48
Reply from 91.121.91.149: bytes=32 time=59ms TTL=48
Reply from 91.121.91.149: bytes=32 time=56ms TTL=48

Ping statistics for 91.121.91.149:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 54ms, Maximum = 59ms, Average = 56ms

C:\Documents and Settings\Andrew>ping ns28189.ovh.net

Pinging ns28189.ovh.net [91.121.92.19] with 32 bytes of data:

Request timed out.
Request timed out.
Request timed out.
Request timed out.

Ping statistics for 91.121.92.19:
Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),
Control-C
^C
C:\Documents and Settings\Andrew>ping ns27361.ovh.net

Pinging ns27361.ovh.net [91.121.85.17] with 32 bytes of data:

Reply from 91.121.85.17: bytes=32 time=60ms TTL=47
Reply from 91.121.85.17: bytes=32 time=58ms TTL=47
Reply from 91.121.85.17: bytes=32 time=55ms TTL=47
Reply from 91.121.85.17: bytes=32 time=58ms TTL=47

Ping statistics for 91.121.85.17:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 55ms, Maximum = 60ms, Average = 57ms
I can access your servers just fine (well, on the one that responds to HTTP I can).

I'd like to help you as much as possible with this problem since OVH don't seem to want to.

kurtis08
14-03-2008, 14:19
i see i never got answered guess you get what you pay for

kurtis08
13-03-2008, 14:06
timeouts within the dc network theres timeout on the clients side as he doesnt allow pings so thats ok but timeouts within the dc this isnt acceptable

kurtis08
13-03-2008, 14:05
heres and example

1 rbx-28-m1.routers.ovh.net (91.121.92.253) 0.990 ms 0.983 ms 1.091 ms
2 rbx-2-6k.routers.ovh.net (213.251.191.130) 0.494 ms * *
3 * * *
4 * * *
5 040G.gsw-1-6k.routers.ovh.net (213.186.32.158) 4.263 ms * *
6 * * *
7 * pos-0-14-0-0-cr01.newyork.ny.ibone.comcast.net (68.86.85.105) 80.700 ms 80.630 ms
8 * * *
9 * te-0-0-0-4-cr01.denver.co.ibone.comcast.net (68.86.84.33) 135.995 ms 135.971 ms
10 te-0-4-0-0-cr01.seattle.wa.ibone.comcast.net (68.86.90.209) 163.966 ms 164.143 ms 164.114 ms
11 * * *
12 * * *
13 te-9-4-ur01.bellevue.wa.seattle.comcast.net (68.86.177.105) 172.410 ms 172.343 ms *
14 68.85.240.250 (68.85.240.250) 172.291 ms 172.289 ms 172.243 ms
15 * * *
16 * * *
17 * * *
18 * * *
19 * * *
20 * * *
21 * * *
22 * * *
23 * * *
24 * * *
25 * * *
26 * * *
27 * * *
28 * * *
29 * * *
30 * * *

kurtis08
13-03-2008, 14:04
The reason there is no traffic as noone can connect to the the servers from america

if you check previous mrtg graphs we normally running at 50mbps and above now we are not as noone can connect the server

i will provide a traceroute shortly

hashbangbinbash
13-03-2008, 12:46
Oh, I see what you mean. A busy server would netstat alot more than 55 lines, so it's effectively at rest.

I'm sure kurtis08 will be along soon with the traceroute you asked for.

oles@ovh.net
13-03-2008, 12:12
hashbangbinbash a écrit:
>
> Please tell me, how does knowing netstat spits out 55 lines
> enlighten?


you didn't answer me. no traceroute.

netstat 55 lines means there is no activity on the server.


hashbangbinbash
13-03-2008, 11:18
Please tell me, how does knowing netstat spits out 55 lines enlighten?

oles@ovh.net
13-03-2008, 08:04
kurtis08 a écrit:
>
> Hi,
>
> i have 2 100mbps and 1gb server with you and they are buffering to hell
> with 10mbps usage.
>
> USA clients cannot connect to them at all i have timeouts within the dc
> routers


you are in SLA mode. So you have full traffic. You have no limitation in
the traffic.

Do you have any traceroute FROM and TO your customers IP ? What is your
IP / your customers's IP ?

The question is: can you customer "create" more traffic that 10Mbps from
a single IP ? The answer is probably not. But I don't see any traffic on
your server:
[root@streaming3 ~]# netstat -tanpu | wc -l
55



kurtis08
13-03-2008, 05:22
Hi,

i have 2 100mbps and 1gb server with you and they are buffering to hell with 10mbps usage.

USA clients cannot connect to them at all i have timeouts within the dc routers

please could you fix the problem server name is

ns28159.ovh.net
ns28189.ovh.net
ns27361.ovh.net

thanks