iSCSI issues post migration from Core to Scale (Bluefin)

[Chowtime7441]

I just completed the migration from the latest version of Core, 13.0-U6.1 to Scale 23.10.2. And while the migration was mostly uneventful, iSCSI was broken. The iscsi initiators (Proxmox 8.1 systems and Ubuntu Server 22.04) could connect to the portal but no LUNs/iSCSI devices were getting attached to the session. In Core my targets started out with a LUN ID of 1 and incremented sequentially and this wasn't an issue.

Eventually through testing I have found that as long as a LUN with the ID of 0 or 1 is present in the scst security group (iSCSI Group Security in the web ui) and any additional iscsi targets in the security group or portal (if no iscsi initiator restrictions are in place for the targets) have their LUN ids numbered sequentially the LUNs will attach to the iSCSI session.

If I'm doing something wrong straight out of the install here please let me know, otherwise I think this might be a bug in the SCST code.

Thanks!

 

[HoneyBadger]

Hey @Chowtime7441 - thanks for pointing this out.

SCALE (Linux) uses SCST which requires LUN 0 to exist on each target; CORE (FreeBSD) uses ctld which I don't believe has that same hard requirement.

Can you report this as a "Report a Bug" issue using the link at the top of the forums, or from the SCALE UI itself at System -> General -> File Ticket and include a debug? Send me the JIRA ticket # and I'll see if I can add some additional context and commentary.

 

[Chowtime7441]

@HoneyBadger Thanks for that information. I'll open a ticket and include the debug a bit later today.I also realized that I mixed up the version names and release numbers for Scale. Sorry about that.

 

[anodos]

https://github.com/SCST-project/scst/pull/223 We're already fixing this issue upstream.