OVHCloud Network Status

Current status
  • Operational
  • Degraded performance
  • Partial Outage
  • Major Outage
  • Under maintenance
FS#4327 — Orange
Incident Report for Network & Infrastructure
We notice difficulties of access from Orange. We are looking for the problem's cause.


Date: 2010-07-01 11:20:15 UTC
The error of beginners.

Date: 2010-07-01 09:49:22 UTC
During the maintenance
We have migrated the ports from gsw-1 to gsw-2
and we forgot to re-announce the /30
on gsw-2

It continues to work because the routes
were announced in BGP.

After the migration of yesterday, it continues to work.
Only for certain networks which are
prefered in gsw-2. Therefore, the traffic
coming from rbx-1 goes on gsw-2
via th1-1 and were null routed by the announce
of for all of the network.

th1-1-6k#sh ip route
Routing entry for
Known via \"static\", distance 200, metric 0 (connected)
Advertised by bgp 16276
Routing Descriptor Blocks:
* directly connected, via Null0
Route metric is 0, traffic share count is 1

gsw-2-6k#conf t
gsw-2-6k(config)#router ospf 16276
gsw-2-6k(config-router)#network area 0

th1-1-6k#sh ip route
Routing entry for
Known via \"ospf 16276\", distance 110, metric 3, type intra area
Last update from on Port-channel5, 00:00:01 ago
Routing Descriptor Blocks:
*, from, 00:00:01 ago, via Port-channel4
Route metric is 3, traffic share count is 1, from, 00:00:01 ago, via Port-channel5
Route metric is 3, traffic share count is 1


Date: 2010-07-01 09:25:14 UTC
It seems that the origin of the problem is from OVH.
We will make a test on /24 in order to find out
the problem.

Date: 2010-07-01 09:01:01 UTC
No feedback with the problems. We close the ticket.

Date: 2010-07-01 08:58:21 UTC
You should have seen a cut off between 7h15 and 7h40 approximately.

The routing towards Orange must be back to normal
for the customers who were concerned.

If you still have problems towards OVH from
Orange, would you please point it out on We
need to have your IP at Orange.

Date: 2010-07-01 08:53:58 UTC
We reactivate the routing via the ports having problems in order to make a diagnosis.

Date: 2010-06-30 21:10:40 UTC
Apparently the routing to pose
problems from Orange side also. We have opened a
ticket at Orange. we wait.

In all cases tomorrow morning we will make tests
with teams, we've been in contact with them
today at the changeover.

Tough tough...

Date: 2010-06-30 17:08:18 UTC
Thanks! and sorry ...

Date: 2010-06-30 17:07:23 UTC
No idea about the origin of the problem.
We do not want to reactivate this link at 18h
to perform tests. We left thus the
traffic to pass only through Telehouse 1.

Tomorrow morning 7:00, we will reactivate the
port and check with Orange teams
Why the traffic doesn't pass correctly
on their network.

If you have been impacted to this problem, thank you
to post to , your IP with
Orange. This will allow us to make necessary
verifications, tomorrow on the good or bad
functioning to your connection.

Date: 2010-06-30 15:43:57 UTC
An idea submitted to Orange: this morning we changed the
configuration from 2x 1x10G BGP with 2 BGP sessions to
2x10G connection with channels port, with a BGP session
May be at the time of the cleaning configuration works
the / 30, which allows us to make the BGP
was deleted at the same time as the / 30 which no longer serves.

Date: 2010-06-30 15:35:11 UTC
for the moment, the traffic passes through 2x10G
we have on Telehouse 1.

Date: 2010-06-30 15:32:28 UTC
OVH has 2 private peering of 2x10G with Orange. One on
Telehouse 1 and the other on Global Switch.

This morning we switched the configuration on
Global Switch to a port channel configuration

Everything went well.

From 16:45, we notice a fell of 25% of the global ovh
traffic to Orange. 50% on the traffic
at Global Switch. There is no change from Telehouse 1

We just cut the BGP session on Global Switch
everything is back.
It is reactivated. Still 50% of the traffic.
It is not a loss of packets. Is a
net interruption for some customers.
at Global Switch.

We conclude with a problem with Orange. We keep
the port down and contact them to see what is
going on.
Posted Jun 30, 2010 - 15:18 UTC