Hi.
We recently purchased a QNAP TS-883XU-RP to use as our mass storage for our company storing the vhdx files for 2 files servers with my HP DL385 Gen8 as the host server. During testing I was able to connect to the NAS using iSCSI and copy 4TB+ of data with no issues. Now it is on site and with iSCSI connecting it to the HP server (Server 2012 R2) and it remains stable until I start to move data. Using robocopy, I am attempting to move our archive data, ~4TB, to the new VM but after a short period of time, the connection drops.
Our HP server has 4 onboard NICs, all are teamed, the team is configured as switch independent and dynamic load balancing. The NAS also has 4 onboard NICs and they to are teamed for 802.3ad dynamic and the switch has also been configured as per this KB article.
This is what originally happened.
- Connected NAS to server using iSCSI and created the VMs.
- 2 new VMs, one will be archive data and the other will be live data. Both have 2 drives, OS and data storage and all are on the QNAP NAS.
- Joined the VM to the domain and fully updated the Servers (both Server 2019). During this period, no issues at all.
- Started the transfer of data from old servers to new VMs using robocopy - both at the same time.
- After about 10GB of transfer, the host server lost connection to the VM and I am unable to RDP to the host VM but I can RDP to the domain controller that is kept on the same host server and from the DC, I can RDP to the host server.
- The host server loses connectivity to the NAS but I can ping and access the NAS web front end with no issues and the performance is fine.
- iSCSI initiator shows the NAS is disconnected and from the host server I am unable to ping the NAS or any device outside of the Hyper-V vSwitch.
- A host server restart brings everything back to normal.
I then found a site that recommended changing the iSCSI settings to use Microsoft iSCSI Initiator and not the default setting and this removed a 113 warning I had in the Server logs and I enabled jumbo frames. I then tested transferring data on just one server and again I got a disconnect from the VM but this time, I remained connected to the host server but one of my DCs became unavailable and I was able to ping the gateway and internet servers (8.8.8.8).
I believe this is an issue/mis-configuration on the vSwitch or the server NICs but I can find any information related to it.
Look forward to your insight and support!
J.