Hi all, I'm just wondering if anyone else has experienced this.
We have a 3-node 2012R2 cluster, all nodes running core. It's fine, the only test it seems to fail is the cluster communication networks all being on the same subnet- if I set them as the same then the cluster test complains, if I set them as separate then the Dell HITKit complains so it's lose-lose.
The cluster will tick along nicely, we can move things around etc. Then, after some time (?), event log errors start popping up. These can be (sorry, the list is long): 1038, 1069, 1126,1127,1129,1135,1137,1146,1155,1205,1254,5120,5142.
We don't necessarily get all of these error and am not sure which ones crop up first, but it seems like the host networking gets...clogged up? That sounds daft, but if we reboot (drain) the hosts then the problem is resolved, and the cluster carries on for however long.
Microsoft have, in the past, suggested settings to switch off (TCP chimney's etc, a bunch of stuff) on each host NIC and the Dell HITKit is installed on anything directly accessing EqualLogic volumes. We patch the hosts, run the Dell SUU CD against them once in a while to keep drivers, firmware up to date etc.
I'd be grateful for any help- like I said, however daft it sounds it just seems like the networking gets clogged up with data after a while so the adapters freeze up.
?