When starting/restarting a container,
aprsend for an IP address fails with "IP is detected on another computer"
[root@vz ~]# vzctl restart 1406 Restart the Container ... Adding IP addresses: 220.127.116.11 arpsend: 18.104.22.168 is detected on another computer: XX:XX:XX:d6:84:66 vz-net_add WARNING: arpsend -c 1 -w 1 -D -e 22.214.171.124 eth0 FAILED ... Starting the Container ...
Another symptoms you may observe for such container:
- Parallels Power Panel is not accessible for this container or drops connection from time to time.
- Container is not pingable/cannot be accessed via SSH.
How to troubleshoot
Check the arp entries for the container in question on the node where it resides:
[root@vz ~]# arp -an | grep 126.96.36.199 ? (188.8.131.52) at XX:XX:XX:d6:84:66 [ether] on eth0 ? (184.108.40.206) at * PERM PUP on eth3 ? (220.127.116.11) at * PERM PUP on eth2 ? (18.104.22.168) at * PERM PUP on eth1 ? (22.214.171.124) at * PERM PUP on eth0
In the example above we see that there is an entry for MAC that does not belong to localhost:
? (126.96.36.199) at XX:XX:XX:d6:84:66 [ether] on eth0
arpingfrom another node in the network, most probably, you will see that reply comes from two different MACs, as on the example below:
[root@vz2 ~]# arping 188.8.131.52 -I eth2 ARPING 184.108.40.206 from 220.127.116.11 eth2 Unicast reply from 18.104.22.168 [XX:XX:XX:D6:84:66] 1.078ms Unicast reply from 22.214.171.124 [YY:YY:YY:00:1C:26] 545.118ms
Network conflicts should be resolved by network administrators.