I had today downtime 5 hours on php-fpm(which means all sites was down since they are all php based)
and i starting to think it is not problem in overload beacuse at the same time other server with apache and whm was down too 5 hours.Since apache died there or it was corrupted,not sure since i restarted it in hurry.I have following configuration ,master domain serves thumbs and mysql database (nginx+phpfm server)where is storing info about clicked thumbs,and child domains are on server with whm and apache.I am not 100 percent sure,but i think every time when php-fpm was down httpd server was down there too.Also i upgraded phpfpm yesterday to 5.2.11 php version but i doubt that have any connection,i mean i dont see how that will cause apache down on another server.But again who knows.Downtime appear again before one hour and it seems solution to fix it was same,restarting apache on child server.
Anyway,any idea why does died/corrupted apache affect php-fpm on external server?I disabled connection to prevent this to happen again.

roguehosting wrote:
> I had today downtime 5 hours on php-fpm(which means all sites was
> down since they are all php based) and i starting to think it is not
> problem in overload beacuse at the same time other server with apache
> and whm was down too 5 hours.Since apache died there or it was
> corrupted,not sure since i restarted it in hurry.I have following
> configuration ,master domain serves thumbs and mysql database
> (nginx+phpfm server)where is storing info about clicked thumbs,and
> child domains are on server with whm and apache.I am not 100 percent
> sure,but i think every time when php-fpm was down httpd server was
> down there too.Also i upgraded phpfpm yesterday to 5.2.11 php version
> but i doubt that have any connection,i mean i dont see how that will
> cause apache down on another server.But again who knows.Downtime
> appear again before one hour and it seems solution to fix it was
> same,restarting apache on child server. Anyway,any idea why does
> died/corrupted apache affect php-fpm on external server?I disabled
> connection to prevent this to happen again.
>
> Posted at Nginx Forum:
> http://forum.nginx.org/read.php?3,18368,18368#msg-18368

You should provide us with code snippets showing your configurations and
also the versions of each of the programs you're using.

If don't know where to start looking for this problem and you're using
mysql-5.1, I'd recommend to disable the query cache and see how it goes.

mysql > SET GLOBAL query_cache_size = 0;

I've had some terrible nightmares because of it.

http://www.mysqlperformanceblog.com/2009/03/19/mysql-random-freezes-could-be-the-query-cache/

-Jason

On Oct 30, 2009, at 11:36 AM, "Juan Fco. Giordana" <juangiordana@gmail.com
> wrote:

>
> roguehosting wrote:
>> I had today downtime 5 hours on php-fpm(which means all sites was
>> down since they are all php based) and i starting to think it is not
>> problem in overload beacuse at the same time other server with apache
>> and whm was down too 5 hours.Since apache died there or it was
>> corrupted,not sure since i restarted it in hurry.I have following
>> configuration ,master domain serves thumbs and mysql database
>> (nginx+phpfm server)where is storing info about clicked thumbs,and
>> child domains are on server with whm and apache.I am not 100 percent
>> sure,but i think every time when php-fpm was down httpd server was
>> down there too.Also i upgraded phpfpm yesterday to 5.2.11 php version
>> but i doubt that have any connection,i mean i dont see how that will
>> cause apache down on another server.But again who knows.Downtime
>> appear again before one hour and it seems solution to fix it was
>> same,restarting apache on child server. Anyway,any idea why does
>> died/corrupted apache affect php-fpm on external server?I disabled
>> connection to prevent this to happen again.
>> Posted at Nginx Forum:
>> http://forum.nginx.org/read.php?3,18368,18368#msg-18368
>
> You should provide us with code snippets showing your configurations
> and also the versions of each of the programs you're using.
>
> If don't know where to start looking for this problem and you're
> using mysql-5.1, I'd recommend to disable the query cache and see
> how it goes.
>
> mysql > SET GLOBAL query_cache_size = 0;
>
> I've had some terrible nightmares because of it.
>
> http://www.mysqlperformanceblog.com/2009/03/19/mysql-random-freezes-could-be-the-query-cache/

Disabling the cache is not a good solution.

Either upgrade the entire stack, or downgrade to a point where the
cache will work.

Try postgresSQL, convert to Innodb or try MariaDB.

In a high performance network I would not disable but increase the
cache.

-Jason

On Oct 30, 2009, at 11:36 AM, "Juan Fco. Giordana" <juangiordana@gmail.com
> wrote:

>
> roguehosting wrote:
>> I had today downtime 5 hours on php-fpm(which means all sites was
>> down since they are all php based) and i starting to think it is not
>> problem in overload beacuse at the same time other server with apache
>> and whm was down too 5 hours.Since apache died there or it was
>> corrupted,not sure since i restarted it in hurry.I have following
>> configuration ,master domain serves thumbs and mysql database
>> (nginx+phpfm server)where is storing info about clicked thumbs,and
>> child domains are on server with whm and apache.I am not 100 percent
>> sure,but i think every time when php-fpm was down httpd server was
>> down there too.Also i upgraded phpfpm yesterday to 5.2.11 php version
>> but i doubt that have any connection,i mean i dont see how that will
>> cause apache down on another server.But again who knows.Downtime
>> appear again before one hour and it seems solution to fix it was
>> same,restarting apache on child server. Anyway,any idea why does
>> died/corrupted apache affect php-fpm on external server?I disabled
>> connection to prevent this to happen again.
>> Posted at Nginx Forum:
>> http://forum.nginx.org/read.php?3,18368,18368#msg-18368
>
> You should provide us with code snippets showing your configurations
> and also the versions of each of the programs you're using.
>
> If don't know where to start looking for this problem and you're
> using mysql-5.1, I'd recommend to disable the query cache and see
> how it goes.
>
> mysql > SET GLOBAL query_cache_size = 0;
>
> I've had some terrible nightmares because of it.
>
> http://www.mysqlperformanceblog.com/2009/03/19/mysql-random-freezes-could-be-the-query-cache/

Well,i disabled mysql access plus i removed cron but this happened again,but luckly this time i puted uptime checker so i was able to react fast.This is most bizzare server problem i ever had.I really dont see how broken apache affect php fpm on another server now when is only static content pulled.I remember before i was testing cheeroke web server,and as test i made it to serve only those thumbs and nothing else,and one day i noticed thumbs gone and cheeroke cpu usage went to 100(probaly some kind of corruption)so since then i use only nginx to serve thumbs.
I will do following steps in attemp to fix this problem:
first convert apache to nginx,and if that not helping then moving thumbs to apache server completely and cut any external connection.