|
> This sounds layer 2ish. Do you have MTU problems between interfaces?
> Try pinging between all the machines with large pings (>935) and see
> what happens. Any CRC errors on the interfaces or switch ports? Any
> Duplexing errors?
I think you've found the root of the problem:
[root@beta root]# ping -I 206.27.7.52 -s 1024 ws4
PING ws4.listingbook.com (206.27.7.8) from 206.27.7.52 : 1024(1052) bytes of
data.
--- ws4.listingbook.com ping statistics ---
3 packets transmitted, 0 received, 100% loss, time 2017ms
This happens on both real servers. The director has no problems with the
large ping:
[root@delta root]# ping -I 206.27.7.52 -s 1024 ws4
PING ws4.listingbook.com (206.27.7.8) from 206.27.7.52 : 1024(1052) bytes of
data.
1032 bytes from ws4.listingbook.com (206.27.7.8): icmp_seq=1 ttl=128
time=0.638 ms
1032 bytes from ws4.listingbook.com (206.27.7.8): icmp_seq=2 ttl=128
time=0.553 ms
--- ws4.listingbook.com ping statistics ---
2 packets transmitted, 2 received, 0% loss, time 1006ms
rtt min/avg/max/mdev = 0.553/0.595/0.638/0.049 ms
------------------
Pinging ws4 on the eth0 interface with large packets works great, it seems
to be only the tunl0:0 device that is incapable of the large ping.
I'm not sure how to look for CRC or Duplexing errors.
Where now?
-Jacob
|
