I know this has been asked many times over, but I can tell you every trick stated doesn't help me and I bet there are other out there still struggling like me to where I want to rip out those Broadcoms and get Intel's. Well not so easy to do when HP DL380 servers come with Broadcom LOM's right?
I have the same issue with all my HP DL380's whether they're in a cluster or a stand alone Host. They have all the HP 1Gb 4-Port 331FLR's or 331T's, which are basically Broadcom's so even the LOM's are. I was back and forth with MS and HP Support Teams only to hit a brick wall and not have a resolve so issue still happens......even today which is why I'm finally writing this.
Did the usual such as VMQ off, PowerSettings Off, updated driver to the "b57nd60a.sys" as noted in MS KB 2986895 with no luck. My VM's themselves only have one NIC connection though now thinking about trying dual NIC's to VM's with issue.
I know it's not clustering or storage as per say and the reason is that I have the issue with a couple of my stand alone hosts that use local drive storage on server. What I do know is that the issue seems to be when the VM is stressing the network for long periods of time as it's only happening on my VM's (running 2012 R2 as well) that are the backup servers. These VM Servers that drop off the network use VEEAM B&R as well as stream Azure Backups. It seems when Azure is uploading for a long period of time this is when the VM's disconnect from the network and I get the yellow triangle of aggravation. In the cluster I can migrate the VM and be back up and running, though sometimes after migrating it'll drop later that day or maybe not. On the stand alone hosts I have no choice other than to reboot the VM to get it back up and running.
Again, I'm just putting this out there because I'm sure I'm not the only one still having this issue even after doing everything people suggest. Please don't just post an answer that's a re-post from another thread or said links as I've been through them all. What I'd like to have is others that still have the same issue as i and can confirm they have a fix such as just put in Intel NIC's. After all I have both a Cluster with HP SAN as well as HP stand alone hosts with local storage and have the exact same issue. Just funny how every other VM on the Host either Cluster or Stand Alone are unaffected when they use the same network paths.
Thanks, Mark