Hello? Mods? TubaDiva?

Been noticing over the last hour or so that the buttons on top of the message boxes (both ‘Quick Reply’ and conventional) no longer work. For any linking, quoting, bolding, italics, etc. … that has to be hand-coded.

EDIT: 502 error upon making this post, but I knew it had gone through.

Again, very bad right now

Yep, also hitting more errors now.

Lemme see if I can get the techs to look at this.

Sorry for your inconvenience. ::: sigh :::

Jenny
your humble TubaDiva
Administrator

Got another 504 at 1:29 CST today. Got a few more a couple of days ago. Even when not getting the timeout messages, usually after the first 2-4 pages load up fairly quick after each click, the rest run much slower often, perhaps 3-10 seconds a page, before sometimes locking up. It’s sporadic, but somewhat has been going like this for a few days. Still a vast improvement over last week. I’m fine with it, just letting you know if it helps.

I wonder if it gets better when people give up and go somewhere else for a while. One long-term issue with this problem is it’s likely to reduce board membership. The hardcore will likely stick around, but the more casual will naturally drift to other forums rather than wait for the pages to load or retry when they get the timeout. Even if someone thinks they’ll come back later, they may eventually spend more and more time hanging out in whatever corner of the internet they end up in during the outage.

From the description about locking in the large database, it sounds like a very tough category of problem to fix with a band-aid. Once the system starts hitting the limits of the ability to scale for performance, typically significant changes are need. Faster disks may help the system process the locks faster, but there’s an obvious cost to that and it would just be kicking the problem down the road a bit.

Nope, not in my case. For the last 90 minutes, got two more time outs and a 502. Things to a crawl on most pages loading. It’s the same ole same ole, running fine for a little while, then bat crazy the rest.

Just got a couple of 502 errors in the past 5 minutes.

Me three. Feb. 26 13:02 Martian Standard Time. (PST.)

About 10 minutes ago things got ridiculously bad for at least 5 minutes.

Back in my day, we fixed things with baling wire. You got baling wire?

I for one am holding you personally responsible.

:smiley:

I’ve had multiple 50x errors and lots of loading pages just hanging there, sometimes for 30 seconds, sometimes for a couple of minutes. Started this morning, still happening right now.

ETA: I posted this by hitting “Post Quick Reply” and then had things hang for over 90 seconds before I got a 502 error.

The main page seems to be down now, too.

I always heard the phrase “fixed with chewing gum and bailing wire.” Maybe they are out of chewing gum this time.

…Not with a bang but with a whimper…

Sent from my SM-N950U using Tapatalk

I’ve got duct tape and c-clamps. ftg has baling wire. TubaDiva, holler if you want us to take a look at it.

Relax, all right? My old man is a television repairman; he’s got this ultimate set of tools. I can fix it.

Anybody got a hairpin?

Maybe the system has the flu. Like everybody else.

Talking to TPTB. It will take a red hot minute or two so please hold on.

Jenny
your humble TubaDiva
Administrator

Generally seems to be okay (not oaky) at 6:30 am CET, with a small hiccup or two. But I don’t post much.

Another possible solution to dealing with a database being locked for too long is to reduce the size of the database. In other words, take the oldest threads and archive them, then remove them from the database and compact it.

By “oldest,” I mean the ones that haven’t been accessed in any way for the longest time. If a thread is actively updated, or even just gets a lot of views, don’t archive that one. That should minimize the impact.

Then just throw up a link to the archives on the front page of the board, maybe calling it the “post graveyard” with a description about how we finally had to permanently put down our beloved zombies.

(I’m assuming the size of the database is part of the problem, since I don’t think we’ve hit the highest activity ever. But we inherently must have hit the largest size the database has ever been. There are, of course, always more posts in the archive than there were yesterday.)