Apache workers hangs as gracefully finishing

[Svippy]

I am running a site with SMF 2.0.7 on Apache 2.4.59, running Debian 11 with a 5.10.209-2 kernel, PHP 7.4.33 against MySQL 14.14.

When I run /usr/sbin/apachectl fullstatus, I notice a lot of workers were stuck in gracefully finishing (G), all from the SMF site.  I eventually configured Apache to time them out after 900 seconds, and while that did solve some of it, under heavier loads, they still persist in this state for too long, and the server is unable to respond.

I have searched through Apache's error.log for some clues, but I cannot seem to figure out why SMF seems often to get stuck gracefully finishing, and there seems to be no pattern either.  It can be the main page, a specific topic or board.  It happens some times (though fortunately a minority of the times), so it's not even consistent.  I generally perplexed.  Any clues as to what to look for?

[Aleksi "Lex" Kilpinen]

To be honest I've got no clue, but wanted to point out that 2.0 is currently version 2.0.19 so you are quite a few versions behind.

[Svippy]

So it appears I had gotten it wrong, it was not 2.0.7, but 2.0.17.  Nevertheless, I decided to attempt to patch upgrade to 2.0.18, and I noticed it failed on the SSI.php file.  Inspecting it, I notice it was still a 1.1.x version.  Eventually, I just ended up upgrading through the upgrade package to 2.0.19, and check the difference, and it looks fine.  I did try to run upgrade.php, but it ran into an unknown error.  What database differences are there between 2.0.17 and 2.0.19?

Regardless, I will now try to observe whether the wrong versioned SSI.php was the one causing the issues.

[Aleksi "Lex" Kilpinen]

If memory serves, there are no changes in DB that you should be worried about - But I would still be curious to know why the upgrade failed, it could even be related to your issues.