Hi there,
Over the last few months we've noticed Wingate has become really unstable with the underlying OS either no longer passing traffic or the box simply just blue screening with a driver error.
This is not just on one box either, it's across three different boxes that we have deployed with very similar setups.
Brief overview of the builds.
2 x Server 2012 R2 running on vmware esxi 6.5 U2 with the E1000E network cards
1 x Server 2016 running on vmware esxi 6.5 U2 with the E1000E network cards
The Server 2016 is a fresh build to try to solve the issue with crashing however this one is worse and blue screens randomly several times a day!
The blue screens are all to do with the network stack and are either:
Driver IRQL not less or equal on driver e1i63x64.sys
Attempted write to readonly memory on driver tcpip.sys
There's nothing special about the setups. Two NIC's one outside with a public address and gateway, one inside with just the LAN ip and mask. They do have a few static routes for reachability of other internal networks, but that's fairly normal.
It all seems to have started in the last round of windows updates that have occurred over the last few months specifically to fix this RDP issue with CredSSL. However that can't be the main cause, it's all to do with the driver stack.
Why are we not using the VMXNET3 driver? Well, I have tried and the reason is performance! I have never been able to get the speed through Wingate and the VMXNET3 driver over around 20Mb down and 180-200Mb up. With the E1000E driver we easily get 800-900Mb each way (1Gb connection to the internet)
For experimentation I have built another Server 2016 box with the VMXNET3 drivers to try the various tweaks that are around but so far, speeds are as above. Shocking! And since building this box a few hours ago and installing a fresh install of Wingate I'm already getting blue screens!
This is a pretty serious problem as the software is now no longer stable and we cannot use it and have been forced to disable content filtering on one site already due to the constant crashing.
Incidentally, one box that the new Server 2016 install replaced has had Wingate un-installed as well as the Wingate Network Driver and it's now completely stable again and has been up since yesterday evening (early days, but before this box was crashing a few times a day!)
Any thoughts on this issue? If you need access to one of these boxes, please say as RDP or direct console access can be arranged.
Thanks in advance.