Comment 204 for bug 2004555

I just realized that for iSCSI based systems only those using "shared targets" are affected, like I mentioned on comment #57. We forgot to mention it in the final errata.

Regarding the multipathing, using multipathing there are additional issues that can lead to leaks, and we expect all production environments to use multipathing.

It could also be that they haven't properly checked it or that their storage system is issuing the "Power-on or device reset Unit Attention event" that prevents the issue from happening, like I observed in an HPE 3PAR FC system.