6120XG stops passing traffic, must be rebooted frequently

[branfarm]

Hi there,

I'm a longtime Cisco networker who is now entering the HP network space and I need help. I recently picked up 2 C7000 enclosures with 2x 6120XG switches in each enclosure, and the 6120's are giving me grief.

I'm working on setting up the first enclosure and I'm having an odd problem with my windows boxes where everything will work fine for awhile, then randomly stop working. And when I say "everything works fine" I mean, I can ping hosts inside this enclosure, and they can ping out. If I reboot the 6120 in slot 1, the network starts working again. It seems to only affect Windows hosts (I have PSP 8.50, with NIC-bonding, NFT-PO configured). I've tried deleting the team, forcing preference, removing tx/rx validation, disabling links on the second 6120 (to isolate one switch), all with out success. The only thing that works for sure is rebooting.

My topology is like this: Dual Cisco 6500 switches with 10Gig modules. The 6120 in slot 1 connects to the primary 6500, and the 6120 in slot 2 connects to the secondary 6500. I have one internal X-link port enabled ( port 23) and the other is disabled. Ultimately I would like to be able to have a square topology, with spanning tree blocking the X-link between the 6120's in normal circumstances.

Any help would be greatly appreciated.

Thanks,

Brandon

[weedtech]

Brandon,

Unfortunately I cannot offer a solution to your issue, however I can say that we have experienced several strange issues with our 6120XG switches that sound similar to yours.

I was wondering what the status is on your issue. Was HP able to resolve it?

Like you we were a Cisco centric shop that implemented these 6120XGs and are now struggling to resolve these strange issues.

Our environment is very similar to yours as well, 2 x 6509 (only 1Gb interfaces) connected to slots 1 and slot 2. We did have a square topology but we had to remove the X links (for reasons I can discuss over PM if you would like).

Let me know what your status is, and perhaps we could arrange a PM conversation to discuss our various issues. It may be mutually beneficial.

Regards,
Patrick

[branfarm]

Hi Patrick,

I was finally able to solve the issue, and I believe it was an incompatibility between Cisco STP implementation and HP STP. I eventually switched my entire site to multiple spanning tree. This solved the issue immediately.

I'd be happy to answer any other questions you might have.

Thanks,

Brandon

[weedtech]

Interesting. We currently run MST as well so there must be something else happening under the hood of the 6120's.

[branfarm]

What IOS/Image are you running on your 6500's/6120's?

[Andrew Bialock]

I had the same issue for anything plugged into a 1 GB UTP SFP. I could reboot the switch or hot plug the SFP and it would start working again. HP had me upgrade to Z.14.26 as they claim a lot of hangs and crashes in the previous version Z.14.22.

Thanks,
-Andy

[Jeroen_D]

We had huge issues with those buggy switches;
Try getting to 14_26 firmware first and then retry.
The fixes in that release and in previous releases are very critical! There is one that makes the switch believe he is STP Root after a reboot (and thus makes your STP recalculating) and another that shows a link as up while it is not sending/receiving traffic.
I guess you hit that bug too.