Is there a way to run "Find errors and repair" via Linux CLI?

[Douglas]

That, or if anyone has any suggestions with how to address the issue (noted below), that'd be awesome!

Getting a gateway timeout (via Cloudflare) when I try to run "Find errors and repair" maintenance script.

DNS is at Cloudflare
System is Debian, running Docker.
NginX for static stuff.
Apache and PHP (latest for both).

[shawnb61]

When it finds fixes to do, it keeps track of those in your session.

If you have lots of fixes needed, it can be bigger than can be stored on your session record, which can create lots of weird issues. 

Sometimes this is fixed just by increasing the field on the session record.

Temporarily using file sessions instead of db sessions will usually work.

[Aleksi "Lex" Kilpinen]

If Cloudflare is giving you the timeout, can you try to access the forum via IP directly? Leave Cloudflare out of it, and see if things work right?
You might need to disable 2FA if you are using it, but other than that it should be possible to do.

[Arantor]

You might even get bitten by the connection between nginx and Apache on that score.

The whole feature really needs rebuilding at this point because on any sizeable forum it's just going to flail hard.

[Douglas]

I think many people are forgetting the sheer size of the forums I manage (11,608,338 Posts in 450,996 Topics by 26,826 Members).

I really need to be able to get Cloudflare and NginX out of the equation by doing this via CLI.

The IP is a shared IP for all sites (14) on that server, so accessing HV forums via IP is not possible.

Thankfully, because I have an awesome hosting partner, I can "jump" into the docker environment to access PHP based commands at the CLI.

Long story short, I have a moderated post topic that doesn't exist. Recounting posts doesn't fix that issue.

BTW, just for the record, I am NOT getting any emails or alerts about replies, even though I've fixed my email and have selected to receive both to this thread.

[shawnb61]

To the best of my knowledge, the answer to the question is no, there is no such feature.

Good idea for a feature request.

In the meanwhile, I would go with temporarily using file based sessions.

Maybe setting aside a maintenance window, to minimize user impact.

If you haven't done it in a while, do it in your test system first.

Are you on InnoDB?  This is the kinda thing I used to see on MyIsam.  Haven't seen it since cutting over to InnoDB (way back in 2.0 days).

[Douglas]

I'm on a mix of MyISAM and InnoDB. Tables that get more read action are set to MyISAM. Tables that get written to more (posts, messages, etc) are InnoDB.

Like I said, I haven't had this issue previously... it's only after I switched to a Docker based system, running Cloudflare for DNS that this started happening. Previously, I was able to run the process without issue.