OVH Community, your new community space.

two ovh servers cannot talk to one another


zimsters
01-07-2009, 09:28
Quote Originally Posted by Myatu
From either 92.23.xx.xx host? If so, the problem would be on the host node (to which the main IP is assigned), otherwise it'd wouldn't reach your guests either.

server 1 has the 92.23 ip as host and the 87.98 ip as guest
server 1 host can ping guest fine. shows internal connectivity is ok.

server 2 is 92.23 ip. it cna't ping either server1 ip nor the 87.98 guest, in fact, ANY of the ips held by server1. server 2 can ping the gateway for server1 and the gateway for server2 just fine. a traceroute dies after it hits the gateway for server2.

Myatu
30-06-2009, 19:40
From either 92.23.xx.xx host? If so, the problem would be on the host node (to which the main IP is assigned), otherwise it'd wouldn't reach your guests either.

zimsters
30-06-2009, 18:13
hey myatu,

i assigned IP 87.98.x.x onto the guest. pings fine from the 92.23 host.

Myatu
29-06-2009, 17:20
Zimster, is it possible for you to assign a failover IP in a different subnet (not 92.23.xx.xx) to one of your servers, and see if that is reachable (ie., ping) from the other server? It does seem odd, but it almost looks like the router doesn't allow a routeback.

zimsters
29-06-2009, 12:43
p.s. neither servers have firewalls running, and allow all incomign /outgoing traffic uninterrupted.

zimsters
29-06-2009, 12:42
marks, think one of the ovh guys has dropped you a message regarding this.

it feels extreme to reformat a production server which is heavily utilized simply because it doens't ping to another server on ovh but works on every single other site on the net. the server pings the gateway fine, pings any other site, pings each others' gateways, but packets beyond the other server's gateway disappears. this suggests that its the gateway blocking the traffic (firewall?) rather than anything to do with my servers.

in comparison it's like reformatting a PC cause the volume is too high, instead of just pressing the volume down button.

marks
23-06-2009, 09:51
if sometimes we suggest a reinstallation is because this way, you rule out that the cause of the problem is any misconfiguration of the system. As you know, we don't support configuring the servers for the customers apart from the installation of the system, and we cannot afford to spend the time going through it to be sure that's not the cause of the problem.

In your particular case, a quick test could be to reboot the systems (both) in Rescue mode and see if they still fail to contact one another. Post here a ping and traceroutes of that, if you don't mind.

That could be a quite a good way to see if it's your configuration problem or if there is really a routing problem.

Thanks

Andy
22-06-2009, 14:23
Contact OVH support via phone/e-mail if you want an official answer. The forum isn't an official support method so there is no guarantee your post will be read by customer support.

zimsters
22-06-2009, 14:21
anyone?

zimsters
19-06-2009, 14:40
Quote Originally Posted by kro
zimsters wrote:
> sounds like a routing issue. both have a default route.
>
> server1: default via 94.23.30.254 dev eth0
> server2: default via 94.23.28.254 dev eth0


check your netmasks too: 255.255.255.0 for the server's IP, 255.255.255.255 for
failover-IPs
--
Felix
OVH Team
these are server ips, all set to right ip/subnet mask. they ping to google fine, and i can ping to them fine from my home pc, so they are online and active, they just refuse to talk to one another

zimsters
19-06-2009, 14:39
Quote Originally Posted by Andy
OVH's response to everything is a diagnostic or format, so don't sound so surprised.

Can you do a tracert from each server to each server and post them here, we might be able to get an idea of the routing issue from that.
traceroute of A->B

traceroute to 94.23.x.x (94.23.x.x), 30 hops max, 40 byte packets
1 * * 94.23.30.254 (94.23.30.254) 0.544 ms <---- A'S GATEWAY
2 * * *
3 * * *
4 * * *

traceroute of B->A

traceroute to 94.23.x.x (94.23.x.x), 30 hops max, 40 byte packets
1 94.23.28.254 (94.23.28.254) 4.340 ms * 0.852 ms <---B'S GATEWAY
2 * * *
3 * * *
4 * * *

kro
19-06-2009, 13:18
zimsters wrote:
> sounds like a routing issue. both have a default route.
>
> server1: default via 94.23.30.254 dev eth0
> server2: default via 94.23.28.254 dev eth0


check your netmasks too: 255.255.255.0 for the server's IP, 255.255.255.255 for
failover-IPs
--
Felix
OVH Team

Andy
19-06-2009, 13:12
OVH's response to everything is a diagnostic or format, so don't sound so surprised.

Can you do a tracert from each server to each server and post them here, we might be able to get an idea of the routing issue from that.

zimsters
19-06-2009, 12:49
this is a really frustrating issue guys.

i have 2 ovh bestof servers. both are for VPSs.

these two servers cannot communicate with one another at all. they used to be able to (i need to scp between them).

no ping, no response or anything. they ping anywhere else fine, and i can ping both from here.

sounds like a routing issue. both have a default route.

server1: default via 94.23.30.254 dev eth0
server2: default via 94.23.28.254 dev eth0


ovh's solution is for me to run the network diagnostics and full server diagnostics with the support partition. otherwise rebuild the server.

WTF, are they serious? these servers work great, they want me to reformat cause there's a routing issue? that's laziness at its best.

any advice much appreciated!

thanks