What do you mean by “This appears to break the Redis integration if you save any additional cache setting afterwards“? You mean first Redis connection shows as “Pass,” but later, if you save any other settings, it shows as “Failed”?
but LiteSpeed Cache removes /home/… from the settings value leaving only /.cagefs/tmp/redis.sock
This is not the bug, and this is correct behaviour. You can check https://docs.litespeedtech.com/lsws/cp/plesk/configuration/#how-it-works
@simplisticsca I double checked the save code and there is no other place where the setting is being changed.
Please give me some details to make sure I understood correctly:
– you edit the settings in WP -> LSC -> Cache -> Object cache, yes?
– The 3 dots(…) in path is used to hide the whole path, yea?
I was able to recreate this issue consistently this past Friday across multiple different sites on the same server stack. During those tests, the /home/{user}/ portion of the socket path was being stripped immediately upon saving. This would validate initially, but then break the Redis integration on any subsequent save because the path was no longer absolute.
Interestingly, as of today, the host setting is remaining intact after saving on those same sites without any changes to the plugin version. To clarify my previous report, the ‘…’ was a placeholder for the Linux username. The actual input was the standard absolute path: /home/{user}/.cagefs/tmp/redis.sock.
@simplisticsca please send a report to Litespeed and share the ID generated. https://docs.litespeedtech.com/lscache/lscwp/troubleshoot/#get-a-report-number
Choose one site that have the consistent behaviour.
Maybe there are some security plugins that run? or some code that changes WP options on save.
