vpsBoard Random Downtime

[Nett]

Has anyone noticed the constant downtime of this forum?

I get 504/502 or timeouts nearly 25% of the time I visit the forum.

 

[wlanboy]

Nope, board is loading fine from my side of the world.

 

[Francisco]

Has anyone noticed the constant downtime of this forum?

I get 504/502 or timeouts nearly 25% of the time I visit the forum.

 

Nope, board is loading fine from my side of the world.

I think PHP sometimes takes a dive and stops serving.

Curtis has put in tickets to us a few times for some pointers and we've given him such.

I don't have a lot of experience with PHP-FPM so I'm not sure if the problem is with

it or what, since we never have this problem with PHP-FCGI.

Francisco

 

[mtwiscool]

This is why i do not use nginx.

Apache ftw.

vary common nginx issue as nginx seams to have issues clearing out old php possess in my tests.

 

[kunnu]

Also I am getting this error, Seems down due to tornado.

 

[Mun]

It might be advantageous to use Apache as the backend PHP processor instead of PHP5-FPM. Apache is MUCH better at processing PHP requests then FPM. 

 

[wlanboy]

Well PHP5-FPM does have some issues on lighttpd too...

 

[Mun]

Well PHP5-FPM does have some issues on lighttpd too...

It has issues period.

 

[splitice]

Its probably just an issue with not enough workers / all workers in use.

Thats 95% (conjectured) of the posts regarding 504/502's on the net.

 

[mtwiscool]

Just use apache if it's set up right it uses about the same ram as nginx.

and it's more reliable.

i switched to Apache and not having to worry if php has crashed is better then getting a million emails saying gateway error.

 

[DomainBop]

I get 504/502 or timeouts nearly 25% of the time I visit the forum.

It has happened to me too.  Next time it happens I'm filing a PayPal dispute and starting "scammer" threads on LET and WHT to let everyone know.  Personal injury lawsuit in the works too because I hurt my finger hitting the back button on the last 504 error.

i switched to Apache and not having to worry if php has crashed is better then getting a million emails saying gateway error.

shutdown -h now works even better

 

[Wintereise]

Just use apache if it's set up right it uses about the same ram as nginx.

and it's more reliable.

i switched to Apache and not having to worry if php has crashed is better then getting a million emails saying gateway error.

Try rm -rf /* for extra stability.

Jokes aside, go away :V

 

[RLT]

I thought it was just me. The board times out a lot and is really slow to load at times.

 

[splitice]

Haha I think its quite fast / reliable.

@manndude: Planning to do some load balancing? Combine it with health checks, two birds one stone

 

[MartinD]

Cost.


I'll have a fiddle with MannDude later when he's around.


When I say fiddle with MannDude, I mean look at tweaking/changing things.


When I say things, I mean php/apache/nginx


You guyyss.

 

[ChrisM]

Cost.


I'll have a fiddle with MannDude later when he's around.


When I say fiddle with MannDude, I mean look at tweaking/changing things.


When I say things, I mean php/apache/nginx


You guyyss.

ohhh boy... lol

 

[HalfEatenPie]

Cost.


I'll have a fiddle with MannDude later when he's around.


When I say fiddle with MannDude, I mean look at tweaking/changing things.


When I say things, I mean php/apache/nginx


You guyyss.

I expected a weiner joke.

 

[drmike]

Cost.


I'll have a fiddle with MannDude later when he's around.


When I say fiddle with MannDude, I mean look at tweaking/changing things.


When I say things, I mean php/apache/nginx


You guyyss.

Well, at least you both are adults. 

But I don't think Curtis plays fiddle or likes fiddling unless it's a dark haired Senorita.

But I know he likes some downright good 'Murica fiddle picking.

 

[drmike]

Has anyone noticed the constant downtime of this forum?

I get 504/502 or timeouts nearly 25% of the time I visit the forum.

Any rhyme or reason to when the problems are occurring?  For instance, seeing such more often around top of the hour? like :00 - :05?  I know there was a matter relating thereto, some hours involving backups/dumps.

PHP like in most places is problematic.  Tuning more workers and other optimizations should be considered. Probably double stacking HTTP daemons overdue also for isolation, cleaning up issues, and general less load on the server doing the heavy lifting, like ---> Nginx ---> Lighty (or whatever currently is serving such).

I am frequent user of the site   Haven't been experiencing ANY of these problems recently, but that doesn't mean you are not.

 

[MannDude]

Backups run every two hours so, at even number hours (EST) it lags for about 30 seconds. So at 00:00, 02:00. 04:00, etc this happens.

Otherwise I've not really noticed any performance issues. Last week there was a day where I had to restart nginx/php a few times, but that (for me) is semi-rare. Maybe once or twice a month I'll notice it myself. 

I'll see what options are available. I've not outgrown the www-server VPS in terms of resources, so I need to see how to better use these resources to stop this from happening.