Please report TPU site issues

[rtwjunkie]

Got a lot of 502's bad gateway today, and the site did not load properly several times.

It was down almost 3 hours. I'm sure W1zzard knows.

 

[P4-630]

It was down almosr 3 hours. I'm sure W1zzard knows.

For me it was not down, but got loads of 502's and the site was not loading properly before.
It seems that the problems are solved, working again as usual.

 

[manofthem]

NextPowerUp is giving me issues, can't load any of the pages. Been getting this error for a couple of hours now:

Error: unable to write file /home/nextpowerup/www/includes/../../smarty/cache/wrt55bd8a8d564042_29608030

 

[natr0n]

I got bad gateway in the morning / afternoon for a while.

 

[micropage7]

I got bad gateway in the morning / afternoon for a while.

i got too but not sure from TPU or just from my crappy connection

 

[BiggieShady]

Nextpowerup can't write to cache when accessing any article:

Error: unable to write file /home/nextpowerup/www/includes/../../smarty/cache/wrt55bdc7a95af9b8_94483004

Homepage works fine though

 

[W1zzard]

We had some kind of creepy disk error which wasn't detected properly. Everything is fixed now

 

[Caring1]

I'm having an issue accessing any page in the forum, they are loading, but extremely slow, only this site is affected.

 

[W1zzard]

I'm having an issue accessing any page in the forum, they are loading, but extremely slow, only this site is affected.

is the main site affected too?

 

[Caring1]

Seems to be running fine now.
Previously I couldn't access the home page either, it was almost like it was going to time out, but it still loaded.

 

[Tatty_Two]

One thing I have noticed that seems to come and go is when I go into my inbox to clear out some old conversations (with no new messages) and I tick the selected conversations I want to leave I get a pop up message to say I have a new conversation message but there is nothing new.

 

[qubit]

I'm having an issue accessing any page in the forum, they are loading, but extremely slow, only this site is affected.

is the main site affected too?

I haven't noticed any performance issues.

@Caring1 Are you accessing the site from work? If so, perhaps the poxy server is getting in the way?

 

[EarthDog]

Run a tracert and see where things go sour on the way...

 

[Aquinus]

I was randomly met with a gateway error earlier in the day during an AJAX request while I was typing a post. It never reappeared.

@W1zzard : You wouldn't happen to use Nginx as a reverse HTTP proxy to other internal web servers, would you? If you are it's worth noting the free version of Nginx only supports HTTP 1.0 on reverse http requests, so the connection with the proxied server is never kept alive (keep-alive was implemented in 1.1). Is it possible that the shear number of ajax requests is causing port exhaustion on whatever is running Xenoforo? Usually Nginx comes back with this error when it has trouble doing the reverse request which could indicate an overloaded web server on the inside with nothing to fail over to. Something like HAProxy can do keep-alive with proxied servers IIRC and might be better suited for the task and may reduce the impact of AJAX heavy environments.

 

[W1zzard]

You wouldn't happen to use Nginx as a reverse HTTP proxy to other internal web servers, would you?

Already using HAProxy as frontend LB, the issue is probably with our clustering software which sometimes doesn't properly detect broken http server instances.

 

[Aquinus]

Already using HAProxy as frontend LB, the issue is probably with our clustering software which sometimes doesn't properly detect broken http server instances.

Nifty. I'm suggesting an instance might not be down, just acting odd on just the right requests that might hit a limit or something. I recently had to deal with something similar at work which is why I bring it up.

 

[Caring1]

I haven't noticed any performance issues.

@Caring1 Are you accessing the site from work? If so, perhaps the poxy server is getting in the way?

I'm at home and all 5 computers run fine online, no slowness except that incidence on this site only, it lasted for just over half an hour roughly.
I'd click on a link and wander off to make a coffee, the site would load by the time I got back, every other site loaded right away. I did get a 502 error twice.

 

[EarthDog]

Unable to post for about an hour... java issue.. but I am sure you already know about that.

 

[Aquinus]

<b>Parse error</b>: syntax error, unexpected ')' in <b>/home/techpowerup/www/forums/library/TPUDetectSpamReg/ModelSpamPrevention.php</b> on line <b>192</b><br />

I think W1zz forgot to check his syntax before pushing into production.

 

[W1zzard]

I think W1zz forgot to check his syntax before pushing into production.

Added some features to our custom spam detection, tested it, but didn't consider that i'm exempt from spam checks ^^

 

[Aquinus]

Added some features to our custom spam detection, tested it, but didn't consider that i'm exempt from spam checks ^^

This should save you some headache. Just run it at the top level directory of your PHP code and it will run "php -l" on every file until an error is encountered. I just call the script "php-l".

#!/bin/bash
find $1 | grep "\(php\|html\)$" | xargs -ifoo php -l foo

Having written a lot of PHP in the past, I've found that this script can be a time saver. You can easily edit it to take in a single argument that is the directory to run it on I updated the script myself, but I suspect this would prevent such things from happening in the future.

 

[95Viper]

Getting 502 Bad Gateway errors nginx/1.8.0
Getting Timeout errors when trying proxy sites to reach TPU.
And, when I try Down For Everyone Or Just Me it says the site is down.

 

[Aquinus]

It feels like a web front-end is only responding part of the time and others it's straight up rejecting the connection. It's almost a 50/50 chance that a page will load. It feels like a server might be overloaded.

 

[jboydgolfer]

It was down almost 3 hours. I'm sure W1zzard knows.

i got the bad gateway since yesterday, around 13+ hours here in Mass

 

[tabascosauz]

I also had the 502 Bad Gateway for basically all day, but that's not a surprise. What was a surprise was Thursday's behaviour; I couldn't post / comment / PM. Every time I had finished writing my post and went to hit the submit button, it gave me an error saying that there was an error in the Javascript console. No details, no nothing. Started around noon that day and ended at about 5PM. Weird.