Lost connection to an ISCSI target in MSSQL Cluster

Question

Hello dear community

I have a concern, as I don't know through which way I can contact the Microsoft directly regarding MSSQL matters.

The following situation:

Three Windows Server 2019 have been installed and configured on a Hyper-V. One server has the function of ISCSI target server and the other two servers for the MSSQL. There are 4 NICS. One NIC for the live migration, the second the domain NIC, the third for the heartbeat and the fourth NIC is for the ISCSI traffic. On the ISCSI server two disks existed for the Quorum and the second one for MSSQL data. But there is a problem with the MSSQL data disk which keeps losing connection at different times. The Quorum disk does not. Both disks were configured via MPIO and are displayed in the ISCSI initiator. As described, the MSSQL data disk keeps losing connection which is fixed by restarting the ISCSI target server.

I'm at the point where I don't know what could be the cause here, since it explicitly affects this disk. The disks are also backed up by Veeam Backup & Replication and could be restored during the restore test. However, this problem existed before the backup was set up.

The error message in MSFC is not helpful, as it only states that the connection to the disk has been lost.

What can I do?

Answer 1

Hello

Thank you for your question and reaching out. I can understand you are having issues related to iSCSI disks.

As you have MPIO which means there should be two or more paths to Storage or NAS for failover in that case you should lose iSCSI disks or connections.

Please check from switch or network side if there is any Drops or Latency in network of Storage connections.

Please also check iSCSI event logs in Event viewer and Storage logs for further clue .

--------------------------------------------------------------------------------------------------------------------------------------

--If the reply is helpful, please Upvote and Accept as answer--