Server Downtime - May 13th-16th, 2025

[LilianaUwU]

Honestly, with all the messaging to the effect of the forum being on borrowed time, I'm pleasantly surprised that it returned at all.  I was starting to wonder how I'd keep up with road-related developments around the country without it.

I got a Discord server.

[froggie]

Alerts don't appear to be working presently.  It's showing I have one, but I cannot read it.

(EDIT) When I open up alerts as a separate page, I can access them.

[Alex]

I'm getting those same 502 errors, about 20% of the time that I click on a topic.

When the forum first came back up, the "users online" and "most online today" seemed to hover at around 300, which seemed reasonable, but I now see in the last hour that the "most online today" spiked again to 28K - so I'm concerned the bots might be trying to come back?

Modern algorithms (misbranded AI) are a cancer.

Also, I'm getting a bunch of 502 errors if I refresh pages too quickly.  I'm taking it that this some form of mitigation against the bots?

The 502's are a result of me trying to correct some errors with the forum. I had to cut short my work on this yesterday as I had plans and didn't want to make everyone wait too much more. Hopefully I will clean up the 502 error problem this morning.

[Alex]

I wasn't surprised the forum was taken offline (there were warnings about it). I was a bit surprised the forum was down for that many days. Hopefully the "AI" scraper bots won't return any time soon.

I was prepared to leave it offline longer...
I needed time to approach the bot problem from another angle, and eventually surmised leveraging Fail2Ban through Plesk to actively track and block problematic IP's. However, I had major problems with restoring the forum due to the persistent 403's generated by nginx and with software depreciation. They are working on the next SMF finally, but until that is released, I am on my own for correcting code problems such as with the open_basedir settings.

FWIW I've also been having to learn on the fly, as nginx/Apache2 were pretty much voids for me in my knowledge wheelhouse. Since I've been running commands through shell extensively on this, I at least have a better grasp on using it and its capabilities.

[Scott5114]

I wasn't surprised the forum was taken offline (there were warnings about it). I was a bit surprised the forum was down for that many days. Hopefully the "AI" scraper bots won't return any time soon.

I was prepared to leave it offline longer...
I needed time to approach the bot problem from another angle, and eventually surmised leveraging Fail2Ban through Plesk to actively track and block problematic IP's. However, I had major problems with restoring the forum due to the persistent 403's generated by nginx and with software depreciation. They are working on the next SMF finally, but until that is released, I am on my own for correcting code problems such as with the open_basedir settings.

FWIW I've also been having to learn on the fly, as nginx/Apache2 were pretty much voids for me in my knowledge wheelhouse. Since I've been running commands through shell extensively on this, I at least have a better grasp on using it and its capabilities.

Glad to hear that you've found what seems to a solution for the time being, and that you're getting more confident with the tools you're working with.

If we run into further bot problems, Anubis seems to be working pretty well on the wiki, so that would be the obvious next thing to try on the main site.

[MATraveler128]

I keep getting a 502 error when I load certain pages. Still some bugs on my end.

[hotdogPi]

I keep getting a 502 error when I load certain pages. Still some bugs on my end.

I'm not sure it has anything to do with the page. I can't confirm, but it seems to be a constant 40% for me, reloading randomizes the "roll" again. If anything, and I can't confirm, loading a particular page for the first time that's current has a higher chance than reloading any page or viewing a page in a thread where the last post is more than a few days old.

If you get an error after posting, the post does go through. It happened with this one. It was a different error message than the standard 502, though.

[Alex]

The 502 errors are due to this error:
2: Unknown: Cannot call session save handler in a recursive manner

Really getting frustrated trying to solve this. 

[ZLoth]

The 502 errors are due to this error:
2: Unknown: Cannot call session save handler in a recursive manner

Really getting frustrated trying to solve this. 

You got this. We are all counting on you. Thank you for your efforts.

[wanderer2575]

I wasn't surprised the forum was taken offline (there were warnings about it). I was a bit surprised the forum was down for that many days. Hopefully the "AI" scraper bots won't return any time soon.

I was pleasantly surprised the forum came back up what I think was quickly.  I imagined it would be down a week or more.

I join the chorus of those thanking Alex for his time and effort in resolving the issue.

[GaryV]

I've gotten a few random 502 errors. If I refresh it usually clears it up. But sometimes it points me to the top of the page rather than the first unread post.

I had seen the reports of the problems, and then when I got the "forbidden" error I thought the forum had been discontinued. I thank Alex and any others who helped for getting us back online.

EDIT: And alerts are working again for me.

[Alex]

I've gotten a few random 502 errors. If I refresh it usually clears it up. But sometimes it points me to the top of the page rather than the first unread post.

I had seen the reports of the problems, and then when I got the "forbidden" error I thought the forum had been discontinued. I thank Alex and any others who helped for getting us back online.

EDIT: And alerts are working again for me.

I set the forum to 403/forbidden as the bots were still hammering the server with GET requests even after I renamed the forum subdirectory so that the URL would not work. The 403 method was a surefire way to block them.

I think I have the 502 error resolved... Working on one last error now that mostly effects the script processing in the background.

[hotdogPi]

I think I have the 502 error resolved... Working on one last error now that mostly effects the script processing in the background.

Confirmed. I haven't gotten any in a few hours.

[fwydriver405]

No 502 errors today for me (but only one Connection Error message) today so far, compared to the last two days.

Minor things, but noticed too that the "Like" button, as well as some information (like personal messages, date joined, last active, etc) about a user on posts are also unavailable for the timebeing, assuming that is linked to current efforts to combat the bot issues on the server.

Thanks to Alex and everyone else working to resolve such issues.

[Alex]

No 502 errors today for me (but only one Connection Error message) today so far, compared to the last two days.

Minor things, but noticed too that the "Like" button, as well as some information (like personal messages, date joined, last active, etc) about a user on posts are also unavailable for the timebeing, assuming that is linked to current efforts to combat the bot issues on the server.

Thanks to Alex and everyone else working to resolve such issues.

If you encountered a 500 error around 2:30 EDT, that was because of a syntax error I made in a file related to the database that took down the server. 

Otherwise I did deactivate the Like button, as it was causing scripting problems. Some of the other aspects may be because I did not have the original files when I did reinstallations. James or Scott probably has a better grasp of what SMF packages were installed or settings were made such as for date joined and last active.

With that said, I have addressed each of the SMF errors reported. There are still times that the Forum hesitates to where a query exceeds 5 seconds. The jam has something to do with the Display.php script from SMF. Since I've been working on this almost nonstop since 7:30 this morning, I am going to say that is an acceptable issue considering that it seems that functionality is vastly restored. That plus that Fail2Ban and a modification to the nginx directives to block all bots seems to be addressing the thousands of bot GET requests from before.

[Alex]

If you got a 502 error just now, that is because I restarted the server. The memory usage was again topped out. Another thing for me to monitor...

[Molandfreak]

I miss the like button 

[hotdogPi]

If you got a 502 error just now, that is because I restarted the server. The memory usage was again topped out. Another thing for me to monitor...

I saw it, but I thought it was because I was testing if forum search was back (it's not) and it was blocking me specifically for high usage.

[hotdogPi]

Idea: What about a server restart every X hours on the hour (X would be chosen based on how long a restart takes) so the outages are expected rather than unexpected?

[JREwing78]

I appreciate the hard work trying to keep the forum up. It's been working just fine for me today.

[vdeane]

as well as some information (like personal messages, date joined, last active, etc) about a user on posts are also unavailable for the timebeing

I thought the left columns with user information looked unusually short... I guess that's why.

[ElishaGOtis]

I miss the like button 

Like

[ElishaGOtis]

Idea: What about a server restart every X hours on the hour (X would be chosen based on how long a restart takes) so the outages are expected rather than unexpected?

Or a daily downtime scheduled? (I.e. forum unavailable between 0800-1000 GMT M-F).

Alternatively, would universal captcha to access the site (often by Cloudflare or similar companies) be possible like on some other forums I've seen? Or does SMF not support that?

[Rothman]

Working fine for me so far; no "NEW" icons is tolerable.

Miss the "like" feature, though, but no huge loss.

[CoreySamson]

Thank you, Alex, and everyone behind this forum! I greatly appreciate everything you have done to help solve the problems.