Hi Everyone.
Right now I'm facing the same issue with a slight variation, not all my hosts got a PSOD but all of them got their respective vmnic disconnected, the platform is running on an IBM Bladecenter H, 6 HS23 blades (all of them with the Emulex UCN Adapter) only two switches for the first two onboard ports (the 10Gbps nics are disconnected all the time). But I've updated the be2net driver 5 days ago.
Now, my post goes a little further than the PSOD itself... with the two mentioned ports I've configured the virtual switch to use both ports but in the port group for the VM's I've selected the load balance policy of "Explicit failover" using vmnic2 as Active and vmnic3 as Standby, and the port group for Management network with the opposite configuration (Explicit failover using vmnic3 as active and vmnic2 as standby). In the chassis I have two IBM Server connectivity module each with one single group using the six external ports with LACP enabled. In the other end of the cables we have two cisco switches with an etherchannel configuration (one for one of the IBM SCM an one for the orher) this configuration works like a charm from 6 p.m. (around) until 8 a.m. (sharp) next morning and its unable to become stable until the next 6 p.m. cycle, one other thing is that if we use this configuration without any etherchannel grouping it works fine.
So any thought on this issue will be highly appreciated!
Thanks in advanced!