We had an issue early this morning when iSCSI issues broke connectivity with a few of our databases (resulting in a SQL Server Error 21). Attempts to DBCC CheckDB did not work, and the only solution was to restart the SQL Service.

Is there a known reason why an iSCSI initiator session would reset itself out of the blue?

Example below from the NetApp syslog. This set of errors was replicated 4 times (once for each SQL server in production). Only one SQL server was noticeably impacted, however.

[san1: iscsi.notice:notice]: ISCSI: iswta, ISID Rule: new connection from same initiator, shutting down old session 7
[san1: iscsi.notice:notice]: ISCSI: iswta, New session from initiator iqn.1991-05.com.microsoft:sql1.example.corp at IP addr 10.xxx.xxx.123