Resolved Problem php-fpm (70007)The timeout specified has expired: AH01075: Er

[sebgonzes]

Hi

I active 200 domains on one server with php-fpm 5.6, but have several problem :

[Mon Oct 19 17:53:31.969289 2015] [proxy_fcgi:error] [pid 10770:tid 140599495710464] (70007)The timeout specified has expired: [client xx.xx.xx.xx:54335] AH01075: Error dispatching request to :
[Mon Oct 19 17:54:19.939367 2015] [proxy_fcgi:error] [pid 11050:tid 140599462139648] (70007)The timeout specified has expired: [client xx.xx.xx.xx:54542] AH01075: Error dispatching request to :

It seem related to an timeout problem but don't know for now, how to solve it...

 

[UFHH01]

Hi sebgonzes,

the default value "Timeout 60" ( 60 seconds ) in your apache - webserver - configuration could be raised at "apache2.conf" or "httpd.conf" ( depending to your system ). Consider to raise it to 180 or 300 seconds.

 

[sebgonzes]

Hello

I have these with "time" variable :

/etc/httpd/conf.d/fcgid.conf: FcgidIdleTimeout 40
/etc/httpd/conf.d/fcgid.conf: FcgidProcessLifeTime 30
/etc/httpd/conf.d/fcgid.conf: FcgidConnectTimeout 30
/etc/httpd/conf.d/fcgid.conf: FcgidIOTimeout 150
/etc/httpd/conf.d/httpd-includes.confroxyTimeout 1800 (this one is an test to see if problem solve, but not solve nothing)

I use apache 2.4.6....

 

[sebgonzes]

No one have similar problem? we solve it changing all domain from php-fpm to fastcgi, but it's not seem the best solution.
We found similar problem in other system :
- https://forum.directadmin.com/showthread.php?t=50172

but solution not work for us...

 

[IgorG]

Have you tried this solution - http://serverfault.com/questions/500467/apache2-proxy-timeout ?

 

[sebgonzes]

I have try to change various timeout value, but not have a successful result... for now, I have changed all domain to fastcgi

 

[Trindex]

I have similar problem too.
I had Plesk 12 and few websites was configured to run nginx + php-fpm.
After upgrade to 12.5 only that few websites run without problem.
All other websites get nginx 502 Bad Gateway if they run with php-fpm. They run fine if changing to fastcgi.
No errors in logs at all.

 

[sebgonzes]

An update of this item... I still have no solution for this problem, I have put all our hosting with fastcgi, but when I test an wordpress website :
- with php 5.6 on fastcgi : 7.7 seg (very bad result)
- with php 5.6 on php-fpm : 2.2s (better result)
- with php 7 on php-fpm : 0.86s (acceptable result)

The problem is that, it's ok with few website with this php mode, but if I put all 300 domains of server, I will reproduce the problem... Noone have an solution for this error??

Thanks

 

[xernus]

The problem is that, it's ok with few website with this php mode, but if I put all 300 domains of server, I will reproduce the problem... Noone have an solution for this error??
Thanks

I am seeing this too. I do not have a solution either. PHP-FPM works great if there are few sites, but as soon as you hit 150+ sites, these problems will start occurring

 

[Kingsley]

Same here, different types of issue. i just paid $5 for plesk panel. Damn

 

[redcat]

Hi was anybody able to fix this? I have a big site with this exact problem and it's very frustrating.
Switching to fastcgi helps but impacts the performance, as it seems opcache gets flushed every 20-30 seconds. With php-fpm the performance is fantastic until it reaches that error and the website stops responding.
Please help! Thanks

 

[sebgonzes]

Hi was anybody able to fix this? I have a big site with this exact problem and it's very frustrating.
Switching to fastcgi helps but impacts the performance, as it seems opcache gets flushed every 20-30 seconds. With php-fpm the performance is fantastic until it reaches that error and the website stops responding.
Please help! Thanks

Hello

I never obtain an solution to this problem, in our case that is a litte diferent as your case (webhosting server), only have problem if a large number of domain have php-fpm, if a few number have this version, it's works.

 

[lvalics]

I am not sure that this issue is only php-fpm issue. I had fastcgi on older server, moved to a new one and FCGI made this problems. Then switched with a tool to php-fpm and solved the issue, but when I finished to move al 300+ domains over php-fpm, the problem started again. I am still investigating the issue, why is happening this, but is not only a php-fpm issue as I see.

 

[sebgonzes]

In our case, on server with more than 300 domains (between 300 and 350), with fastcgi, all work (but we have a few problem of TTFB with some domains), but sadly, not the same case with php-fpm....

 

[sebgonzes]

To plesk team,
Is this kb the temporal solution of the problem https://kb.plesk.com/en/128227 (to do to each php-fpm version)?

 

[Paddy]

I had this error when I copied a site from a different server. My site was still trying to connect to a mysql database on the old server and the connection was blocked. Once I corrected the connection string, all was fine.

Hope this helps someone.

 

[Jonathan H]

While I hate to resurrect an old thread, I didn't see a clear answer here and this thread is still among the first Google results for this issue.

For those using PHP-FPM, there are two different timeouts that are configured within Apache. There is the plain "TimeOut" setting, and also the "ProxyTimeout" setting, mentioned by others. In order to avoid this error, both Apache and the proxy call to PHP-FPM need to keep the connection going for the required amount of time. The default value for TimeOut is 60, and the default value for ProxyTimeout is to match the TimeOut value.

So for example, if you wanted to double the timeout to 2 minutes, you could throw this into your httpd.conf (and then restart Apache):
TimeOut 120
ProxyTimeout 120

Just bear in mind that changing this means you're willing to keep a request going for up to that long, so make sure you have the resources for it!

 

[Kingsley]

While I hate to resurrect an old thread, I didn't see a clear answer here and this thread is still among the first Google results for this issue.

For those using PHP-FPM, there are two different timeouts that are configured within Apache. There is the plain "TimeOut" setting, and also the "ProxyTimeout" setting, mentioned by others. In order to avoid this error, both Apache and the proxy call to PHP-FPM need to keep the connection going for the required amount of time. The default value for TimeOut is 60, and the default value for ProxyTimeout is to match the TimeOut value.

So for example, if you wanted to double the timeout to 2 minutes, you could throw this into your httpd.conf (and then restart Apache):
TimeOut 120
ProxyTimeout 120

Just bear in mind that changing this means you're willing to keep a request going for up to that long, so make sure you have the resources for it!

How to do this?

 

[Balanca]

While I hate to resurrect an old thread, I didn't see a clear answer here and this thread is still among the first Google results for this issue.

For those using PHP-FPM, there are two different timeouts that are configured within Apache. There is the plain "TimeOut" setting, and also the "ProxyTimeout" setting, mentioned by others. In order to avoid this error, both Apache and the proxy call to PHP-FPM need to keep the connection going for the required amount of time. The default value for TimeOut is 60, and the default value for ProxyTimeout is to match the TimeOut value.

So for example, if you wanted to double the timeout to 2 minutes, you could throw this into your httpd.conf (and then restart Apache):
TimeOut 120
ProxyTimeout 120

Just bear in mind that changing this means you're willing to keep a request going for up to that long, so make sure you have the resources for it!

Thank you Jonathan.. it works

 

[AaronMcH]

While I hate to resurrect an old thread, I didn't see a clear answer here and this thread is still among the first Google results for this issue.

For those using PHP-FPM, there are two different timeouts that are configured within Apache. There is the plain "TimeOut" setting, and also the "ProxyTimeout" setting, mentioned by others. In order to avoid this error, both Apache and the proxy call to PHP-FPM need to keep the connection going for the required amount of time. The default value for TimeOut is 60, and the default value for ProxyTimeout is to match the TimeOut value.

So for example, if you wanted to double the timeout to 2 minutes, you could throw this into your httpd.conf (and then restart Apache):
TimeOut 120
ProxyTimeout 120

Just bear in mind that changing this means you're willing to keep a request going for up to that long, so make sure you have the resources for it!

Thanks for the info, any idea if this can go in the "Additional Apache directives" per domain in Plesk, or does it have to be place in the global config?

In our case this issue doesn't seem to be consistent, it happened earlier today but only for a few hours and only on one domain, then soon after everything went back to normal, but it seems to be an intermittent issue.

Thanks
Aaron