VMware Cloud Community
w3rd
Contributor
Contributor

Host Failure, Crazy Vmnic Allocation issue??

Hey guys, i currently am managing 2 hpdl580s running esx 3.5 v2 in a cluster running only ha / drs... vmotion hasnt been enabled due to networking requirments not being defined. Each of these hosts have a total of 10 ports, 2 onboard, 2 pcix quad port intel e1000. there was a single cable plugged to each host, one subnet / vlan. we issued a cable change for each of the hosts.. the cables were rearranged while the systems were on and no notice was given to the admins that the cable change was going to take place when it did. we didnt get a chance to change the configuration status to maintenance mode, so there was an issue when the cables were plugged back in... one system went compltely offline to virtual center, and the other went into an intermittant connection status with vc. i was able to console in to the box via ilo and see that each of the systems had all 10 of the vmnics attached to the only vswif that was created, vswif0 which is by default needed for network to service console via ip. I was able to remove the vmnics 1 by 1 and eventually the network was isolated to one physical connection.This essentially got me back online and provided the oppurtunity for me to connect the servers via vc again.

I was told this particular connection issue was cause by each of the systems having multiple broadcasts on the same vswif causing broadcast conflicts from the l3 and therefore causing the drop and that this is a known issue in regard to vi3 and this particular scenario. i would like to confirm this if possible.

Reply
0 Kudos
0 Replies