Recent server timeout errors — anyone else getting them?

Yeah doing a record amount of timeouts for me right now.

Same ol’ Same ol’ …even worse than normal the last hour or so.

Seconded.

I don’t blame our humble administrator, and I understand there can be more than one cause, but I would think a serious troubleshooting effort by a capable IT person/team would’ve been able to diagnose the problem by now.

Let’s see, I think the conversation would go like this …

Unknown person that TubaDiva convinced that the board needs help, “We need to bring in a good server consultant for the day to fix the SDMB problems.”

IT manager, “Well that would be 8 hours at $150 per hour. Is it really that bad?”

Unknown person, “It’s pretty bad.”

It Manager, “OK, make the announcement, we’ll be shutting down the SDMB.”
The question is can they find someone to fix the problem and is it worth it to the current owners?
Other possibility, the guy that runs the servers, Googles the problem and does a database re-indexing and tweaks some settings and sees if that fixes the problem. Rinse and repeat until problem resolved. Hopefully this is getting done as I am afraid the IT manager may not be willing to spend his budget on a consultant. He has no attachment to the SDMB.

If the problem happens at 2 AM when use is very low that points to a database issue. If it does not happen then it is probably related to the amount of use. Sometimes systems don’t go bad slowly, they hit a point and it goes bad very fast. Saw that with a network, we had to buy new network hardware because the system suddenly was way too slow.

A consultant would probably charge for 1 day of work at a minimum but if they find the problem fast they may only charge 1/2 day. $150/hour sounds about right but some may charge a little less. Don’t go with the low bidder :slight_smile:

The “Frequently Visited” icons in Safari on my iPad now include “502 Bad Gateway.” Not really a good sign …

I got two of them just trying to open the his thread to report that I am also getting a lot of them. FYI, can’t say I am mad or blaming anyone.

a network issue might be pretty easy to fix , just buy a bigger/faster router or switch. Not sure about the cost , depends on the size needed.

I knew no good would come from allowing avatars!

It’s not about avatars.

It appears to be about the fact that we have 22 million posts and the server is not keeping up with them.

We have been here before when we had smaller servers. We just got bigger servers. Apparently that is not the fix this time.

We’re looking at possible solutions. This is proving to be more difficult than you might think.

Please hang in with us.

Jenny
your humble TubaDiva
Administrator

most of the problems I see are when I add a post. That seems like database inserts are part of the problem. Maybe the indexes are the problem? If the server is big/fast enough 22 million is not a big issue. Could be the database software is old and has not been upgraded or patched. Or as I said above the DB needs tuning.

A lot of stuff is being looked at here both with TPTB and the people at DesertNet, where the servers live. This is not something that can be fixed overnight but please know that they are working on it.

Everyone is aware at how annoying and frustrating this is. It is likewise so for them, they want you to have as good an experience as possible and this kind of stuff is hard to take.

The investigation continues. I’ll keep you posted on what I find. In the meantime, for insurance purposes you might consider making your posts in a file or on another tabbed SDMB page and cutting and pasting it into your thread. That way if the system chokes on your entry you still have it to try to resend. Time consuming and annoying, I know, but I’d rather have an extra page than compose a great post and find it vanished in the ether.

Jenny
your humble TubaDiva
Administrator

I’m doing this sometimes. But just as easily, before I click on Submit, I type Control-A or Command-A (or your iPhone equivalent) to highlight the entire message entry box and copy that to the clipboard. It’s generally safe enough there in case the submission stalls.

I also found that, even easier, if a submission stalls, I can simply click on the Back button to get back to the message entry box, and usually the text will still be there. Your Browser May Vary.

also I see problems almost every time I edit a post. It takes 30 seconds or more to update the post.

I’ve found that a post always Sends, but the screen doesn’t refresh to show it. If you send a post, just wait for the time-our error, hit the back-button, and F5 to refresh. Eventually you’ll see your post. But, before posting, it won’t hurt to Select-Copy, it will then be saved in Notepad, even if not pasted there (until you Copy something else). I don’t thonk I’ve ever had a post lost after sending.

Nitpick: If we are talking about a Windows machine, the Copy function (Ctrl/C) puts the selected data into the “clipboard,” (an invisible location) not Notepad. Of course it can be subsequently pasted into any word processor, but it does not automatically get put into Notepad.

Yes, selecting all (Ctrl/A) and copying your post to the clipboard before hitting “Submit Reply” is a good idea, and a very simple way to avoid Timeout Errors of the First Kind.

Jenny, thanks for the updates. I know you often don’t have anything new to report, but it sure is good to get your take!

I’m adding Gateway 504 to my must-see places. It’s always giving me a time-out. Maybe if I go there someday, I’ll find out what I’m being punished for.

I did lose one post a day or so ago, but that has not been the usual case. I thonk the hamsters for that.

If it helps track down the problem, I get a timeout just about every time I edit a post or report one.

Not a DB expert but to me all the evidence points to a database issue .