You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@pagespeed.apache.org by GitBox <gi...@apache.org> on 2020/07/29 17:52:55 UTC
[GitHub] [incubator-pagespeed-mod] 7starsone opened a new issue #2018: Pagespeed and MaxRequestWorkers
7starsone opened a new issue #2018:
URL: https://github.com/apache/incubator-pagespeed-mod/issues/2018
Hello,
CentOS 7, Apache 2.4.41
it seems that if you set a large pagespeed cache eg. 1 GB and use HTTP connections,
Apache stops responding after something like
`[Wed Jul 29 07:20:17.872755 2020] [mpm_event:error] [pid 16331:tid 140030349564096] AH00484: server reached MaxRequestWorkers setting, consider raising the MaxRequestWorkers setting`
then, it does nothing, ports 80 and 443 are not working anymore even if no high RAM and CPU usage, they are normal, you can just admin the server by web.
Never happened before something like this. I write "it seems..." because I didn't try with the default cache or with other options yet...
And also, how to avoid HTTP connections at all? Is it possible?
I see the _**LoadFromFile**_ directive:
example, if my domain is **https://mydomain.com** and my CDN is **https://static.mydomain.com** and my virtual hosts are in /home/myuser/public_html
Do I need to write
```
ModPagespeedLoadFromFile "https://mydomain.com" "/home/myuser/public_html/"
ModPagespeedLoadFromFile "https://static.mydomain.com" "/home/myuser/public_html/"
```
is that correct?
thanks for your help
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-pagespeed-mod] 7starsone commented on issue #2018: Pagespeed and MaxRequestWorkers
Posted by GitBox <gi...@apache.org>.
7starsone commented on issue #2018:
URL: https://github.com/apache/incubator-pagespeed-mod/issues/2018#issuecomment-669224760
Hello, I reopen the issue because if I don’t set `MaxConnectionsPerChild 0` or leave it as default httpd becomes unresponsive even if not high CPU or RAM usage, and after a while A`H03490: scoreboard is full, not at MaxRequestWorkers.Increase ServerLimit.` on error log. Apache devs say it could be caused by an additional module (so even modpagespeed). Hard to say what, unless you let me know how to inspect what prevent httpd from correctly recycling children. Thanks for now
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-pagespeed-mod] Lofesa commented on issue #2018: Pagespeed and MaxRequestWorkers
Posted by GitBox <gi...@apache.org>.
Lofesa commented on issue #2018:
URL: https://github.com/apache/incubator-pagespeed-mod/issues/2018#issuecomment-669388336
Whats happend if you disable pagespeed?
The error may occur when all server processes go to "gracefull restart" state, and this takes to much time.
What happend if you change MPM EVENt to MPM PREFORK?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-pagespeed-mod] Lofesa edited a comment on issue #2018: Pagespeed and MaxRequestWorkers
Posted by GitBox <gi...@apache.org>.
Lofesa edited a comment on issue #2018:
URL: https://github.com/apache/incubator-pagespeed-mod/issues/2018#issuecomment-669388336
Whats happend if you disable pagespeed?
The error may occur when all server processes go to "gracefull restart" state, and this takes to much time.
What happend if you change MPM EVENt to MPM PREFORK?
Maybe is related to a bug ina apache:
https://www.claudiokuenzler.com/blog/948/apache-2.4-mpm-event-bug-freezing-up-scoreboard-full-after-reload
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-pagespeed-mod] 7starsone edited a comment on issue #2018: Pagespeed and MaxRequestWorkers
Posted by GitBox <gi...@apache.org>.
7starsone edited a comment on issue #2018:
URL: https://github.com/apache/incubator-pagespeed-mod/issues/2018#issuecomment-668642493
Nevermind, it wasn't pagespeed to let Apache reach `MaxRequestWorkers` but the logrotation, which was causing something like
`[pid 21367] poll([{fd=25, events=POLLIN}], 1, 3000) = 0 (Timeout)` in strace and after a while it became unresponsive. With a well configured server I don't think pagespeed may involve something like this. And also, for your info, I noted issues when `ProxyErrorOverride On` and Custom Error pages, the images were not rewritten because the response was not always 200 and `fetch...failed`. Thanks for your help 😄
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-pagespeed-mod] Lofesa commented on issue #2018: Pagespeed and MaxRequestWorkers
Posted by GitBox <gi...@apache.org>.
Lofesa commented on issue #2018:
URL: https://github.com/apache/incubator-pagespeed-mod/issues/2018#issuecomment-667448331
What value have the MaxRequestWorkers?
According to the [apache doc](url) the MaxRequestWorkers directive sets the limit on the number of simultaneous requests that will be served.
In other hand, yes, ModPagespeedLoadFromFile syntax is ok.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-pagespeed-mod] 7starsone edited a comment on issue #2018: Pagespeed and MaxRequestWorkers
Posted by GitBox <gi...@apache.org>.
7starsone edited a comment on issue #2018:
URL: https://github.com/apache/incubator-pagespeed-mod/issues/2018#issuecomment-668642493
Nevermind, it wasn't pagespeed to let Apache reach `MaxRequestWorkers` but the logrotation, which was causing something like
`[pid 21367] poll([{fd=25, events=POLLIN}], 1, 3000) = 0 (Timeout)` in strace and after a while it became unresponsive. With a well configured server I don't think pagespeed may involve something like this. And also, for your info, I noted issues when ProxyErrorOverride On and Custom Error pages, the images were not rewritten because the response was not always 200. Thanks for your help 😄
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-pagespeed-mod] 7starsone edited a comment on issue #2018: Pagespeed and MaxRequestWorkers
Posted by GitBox <gi...@apache.org>.
7starsone edited a comment on issue #2018:
URL: https://github.com/apache/incubator-pagespeed-mod/issues/2018#issuecomment-669224760
Hello, I reopen the issue because if I don’t set `MaxConnectionsPerChild 0` or leave it as default httpd becomes unresponsive even if not high CPU or RAM usage, and after a while `AH03490: scoreboard is full, not at MaxRequestWorkers.Increase ServerLimit.` on error log. Apache devs say it could be caused by an additional module (so even modpagespeed). Hard to say what, unless you let me know how to inspect what prevent httpd from correctly recycling children. By this feedback, anyway, you could do further investigation. _httpd 2.4.43, mod-pagespeed-beta-1.13.35.2-0.x86_64 on CentOS 7_. Thanks for now
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-pagespeed-mod] 7starsone closed issue #2018: Pagespeed and MaxRequestWorkers
Posted by GitBox <gi...@apache.org>.
7starsone closed issue #2018:
URL: https://github.com/apache/incubator-pagespeed-mod/issues/2018
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-pagespeed-mod] 7starsone edited a comment on issue #2018: Pagespeed and MaxRequestWorkers
Posted by GitBox <gi...@apache.org>.
7starsone edited a comment on issue #2018:
URL: https://github.com/apache/incubator-pagespeed-mod/issues/2018#issuecomment-669224760
Hello, I reopen the issue because if I don’t set `MaxConnectionsPerChild 0` or leave it as default httpd becomes unresponsive even if not high CPU or RAM usage, and after a while `AH03490: scoreboard is full, not at MaxRequestWorkers.Increase ServerLimit.` on error log. Apache devs say it could be caused by an additional module (so even modpagespeed). Hard to say what, unless you let me know how to inspect what prevent httpd from correctly recycling children. Thanks for now
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-pagespeed-mod] 7starsone commented on issue #2018: Pagespeed and MaxRequestWorkers
Posted by GitBox <gi...@apache.org>.
7starsone commented on issue #2018:
URL: https://github.com/apache/incubator-pagespeed-mod/issues/2018#issuecomment-667583485
> What value have the MaxRequestWorkers?
> According to the [apache doc](url) the MaxRequestWorkers directive sets the limit on the number of simultaneous requests that will be served.
> In other hand, yes, ModPagespeedLoadFromFile syntax is ok.
Thanks for your reply, it happened two times in the early morning until now, it's like the module fetches more when server resources are underused (and this is good and expected) but this implies the Apache reaches that limit (it was 1024 for MaxRequestWorkers). I increased the limit but I will try loading from file, instead, manually set the `ModPagespeedLoadFromFileCacheTtlMs`.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-pagespeed-mod] Lofesa commented on issue #2018: Pagespeed and MaxRequestWorkers
Posted by GitBox <gi...@apache.org>.
Lofesa commented on issue #2018:
URL: https://github.com/apache/incubator-pagespeed-mod/issues/2018#issuecomment-667657675
The ModPagespeedLoadFromFileCacheTtlMs:
When pagespeed fetches resources via http request, the response have http headers, but when it read from filesystem, no http headers are set. Pagespeed relies in the header cache-control from the resource to set their ttl in the pagespeed cache, but a resource loaded from file don´t have it so the default ttl in the cache is applyed ( 5 minutes)
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-pagespeed-mod] 7starsone edited a comment on issue #2018: Pagespeed and MaxRequestWorkers
Posted by GitBox <gi...@apache.org>.
7starsone edited a comment on issue #2018:
URL: https://github.com/apache/incubator-pagespeed-mod/issues/2018#issuecomment-669224760
Hello, I reopen the issue because if I don’t set `MaxConnectionsPerChild 0` or leave it as default httpd becomes unresponsive even if not high CPU or RAM usage, and after a while `AH03490: scoreboard is full, not at MaxRequestWorkers.Increase ServerLimit.` on error log. Apache devs say it could be caused by an additional module (so even modpagespeed). Hard to say what, unless you let me know how to inspect what prevent httpd from correctly recycling children. By this feedback, anyway, you could do further investigation. httpd 2.4.43. Thanks for now
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-pagespeed-mod] 7starsone edited a comment on issue #2018: Pagespeed and MaxRequestWorkers
Posted by GitBox <gi...@apache.org>.
7starsone edited a comment on issue #2018:
URL: https://github.com/apache/incubator-pagespeed-mod/issues/2018#issuecomment-667583485
> What value have the MaxRequestWorkers?
> According to the [apache doc](url) the MaxRequestWorkers directive sets the limit on the number of simultaneous requests that will be served.
> In other hand, yes, ModPagespeedLoadFromFile syntax is ok.
Thanks for your reply, it happened two times in the early morning until now, it's like the module fetches more when server resources are underused (and this is good and expected) but this implies the Apache reaches that limit (it was 1024 for MaxRequestWorkers). I increased the limit but I will try loading from file, instead, manually setting the `ModPagespeedLoadFromFileCacheTtlMs`.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-pagespeed-mod] 7starsone commented on issue #2018: Pagespeed and MaxRequestWorkers
Posted by GitBox <gi...@apache.org>.
7starsone commented on issue #2018:
URL: https://github.com/apache/incubator-pagespeed-mod/issues/2018#issuecomment-668642493
Nevermind, it wasn't pagespeed to let Apache reach `MaxRequestWorkers` but the logrotation, which was causing something like
`[pid 21367] poll([{fd=25, events=POLLIN}], 1, 3000) = 0 (Timeout)` and after a while it became unresponsive. With a well configured server I don't think pagespeed may involve something like this. And also, for your info, I noted issues when ProxyErrorOverride On and Custom Error pages, the images were not rewritten because the response was not always 200. Thanks for your help 😄
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org