Flex 10 link test failed with strange results.

[philmurray]

Hi, I was all set to release my VC_Flex config to my customer when they asked to just check while I was onsite that smart link was working correctly, for the ESX4 hosts. I removed interconnect 2's cx4 cable and expected everything to be fine, VCM reported all the networks associated with that shared uplink going off line, but the nic's associated to the ESX profile reported differently to my expectation of 2,4,6,8 all being down. Instead the VC for bay 2- ESX profile reported; Lom1a bay1 up, Lom2a bay2 down, Lom1b bay1 down, Lom2b bay 2 up, Lom1c bay1 up Lom2c bay2 down I repeated the reverse of this action, reconnected cx4 interconnect 2, and removed cx4 interconnect 1 and received the opposite equally wrong report. Lom1a bay1 down, Lom2a bay2 up, Lom1b bay1 up, Lom2b bay 2 down, Lom1c bay1 down Lom2c bay2 up I was rather supprised by this, and it had me checking and rechecking the rest of my config for an age before I realised I potentially have a serious issue. I have found links to the vMWare site regarding the apparent failure of smart link(link failure), and have past the relevant beacon probing information to the guys building the ESX enviroment. I am running VC_flex_10_modules, cx4 linked 210, firmware, VC 2.52 OA 490g6 all firmware up to date. Any suggestions ideas would be appreciated. Thanks Phil

[Toni Maatta]

hmm.. I guess I had a similar problem. Updating to VC v 2.12 helped.

[philmurray]

Thanks Toni, that was my next action anyway. Phil