The defaults for php-fpm. So none match and are set lower in PHP-FPM.
Except of start servers
pm.start_servers = 5
pm.min_spare_servers = 5
pm.max_spare_servers = 35
pm.max_children = 50
What would this match to in Apache?
? -> pm.max_requests = 0
These do not seem to have a matching setting in php-fpm from what I have
seen
ServerLimit 16
ThreadsPerChild 16
MaxConnectionsPerChild 0
Not sure how prevalent Apache/Event/PHP-FPM is, but a table correlating
the Apache config settings and PHP-FPM config settings would be useful. At
least for initial setup.
So, is this accurate ...
StartServers ? -> pm.start_servers = ?
MinSpareThreads ? -> pm.min_spare_servers = ?
MaxSpareThreads ? -> pm.max_spare_servers = ?
MaxRequestWorkers ? -> pm.max_children = ?
No Apache equivalent -> pm.max_requests = ?
ServerLimit ? -> no PHP-FPM equivalent
ThreadsPerChild ? -> no PHP-FPM equivalent
MaxConnectionsPerChild ? -> no PHP-FPM equivalent
On 1/14/16 3:24 PM, "Rich Bowen" <rbo...@rcbowen.com> wrote:
>
>
>On 01/12/2016 04:21 PM, Rose, John B wrote:
>> So for example using the default "event" mpm values
>>
>>(http://svn.apache.org/repos/asf/httpd/httpd/trunk/docs/conf/extra/httpd-
>>mp
>> m.conf.in) below in Apache then PHP-FPM would be the same ...
>>
>> StartServers 3 -> pm.start_servers = 3
>> MinSpareThreads 75 -> pm.min_spare_servers = 75
>> MaxSpareThreads 250 -> pm.max_spare_servers = 250
>> MaxRequestWorkers 400 -> pm.max_requests = 400
>>
>> I am not seeing any PHP-FPM equiv settings for ...
>>
>> ThreadsPerChild 25
>>
>> MaxConnectionsPerChild 0
>>
>>
>>
>> Correct?
>>
>
>
>Seems reasonable. What was it set to before?
>
>
>>
>> On 1/12/16 3:41 PM, "Rich Bowen" <rbo...@rcbowen.com> wrote:
>>
>>>
>>>
>>> On 01/12/2016 02:26 PM, Rose, John B wrote:
>>>> Is there a guide to properly matching your apache and PHP-FPM configs?
>>>>
>>>> Looking here Š
>>>>
>>>> https://wiki.apache.org/httpd/PHP-FPM
>>>>
>>>>
>>> I am not seeing that information
>>>>
>>>
>>> php-fpm is documented in the PHP docs at
>>> http://php.net/manual/en/install.fpm.php I'm far from an expert on it,
>>> but the rule would be that if you're proxing from A to B, then B needs
>>> to be able to handle as much traffic as A is forwarding to it. So I'd
>>> expect that what you want is at least as many servers/workers/threads
>>>on
>>> the php-fpm end as you are likely to proxy from the httpd end.
>>>
>>> --Rich
>>>
>>>> On 1/12/16 2:03 PM, "Rich Bowen" <rbo...@rcbowen.com> wrote:
>>>>
>>>>>
>>>>>
>>>>> On 01/12/2016 01:58 PM, Rose, John B wrote:
>>>>>> Can you elaborate on this a bit?
>>>>>>
>>>>>> "..the server pool is smaller than Apache's server pool, causing
>>>>>> too much thrashing. (See the pm.min_spare_servers and so on)"
>>>>>
>>>>>
>>>>> Sure. The way that php-fpm works is that there's a php-fpm daemon,
>>>>>and
>>>>> requests for whatever.php are proxypass'ed over to it. So there must
>>>>>be
>>>>> at least as many threads over on the php-fpm side as you have active
>>>>> over on the httpd side, or there will be thrash in creating those
>>>>> threads when the time comes.
>>>>>
>>>>> So basically you need to line up the php-fpm config with your httpd
>>>>> config, or at least have more threads on the fpm side.
>>>>>
>>>>> pm.min_spare_servers (and max_spare and start_servers) are php-fpm
>>>>> configurations, which you'll find somewhere in /etc/php-fpm.d or
>>>>> whatever your particular distro calls it.
>>>>>
>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> On 1/12/16 12:20 PM, "Rich Bowen" <rbo...@rcbowen.com> wrote:
>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> On 01/11/2016 04:51 PM, Rose, John B wrote:
>>>>>>>> After switching to mod_php from php-fpm we are told the site is
>>>>>>>> working
>>>>>>>> 4x faster using mod_php instead of php-fpm.
>>>>>>>>
>>>>>>>> Any explanation?
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> The cynic in me things that this is a case of old dogs being
>>>>>>> unwilling
>>>>>>> to learn new tricks.
>>>>>>>
>>>>>>> Other things that come to mind is that fpm is grossly
>>>>>>>misconfigured -
>>>>>>> perhaps they allocated insufficient memory to it? Or are running it
>>>>>>> on
>>>>>>> another machine that has a faulty network connection between them?
>>>>>>>Or
>>>>>>> possibly the server pool is smaller than Apache's server pool,
>>>>>>> causing
>>>>>>> too much thrashing. (See the pm.min_spare_servers and so on)
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>> Thanks
>>>>>>>>
>>>>>>>> From: William A Rowe Jr <wr...@rowe-clan.net
>>>>>>>> <mailto:wr...@rowe-clan.net>>
>>>>>>>> Reply-To: "users@httpd.apache.org <mailto:users@httpd.apache.org>"
>>>>>>>> <users@httpd.apache.org <mailto:users@httpd.apache.org>>
>>>>>>>> Date: Saturday, January 9, 2016 1:58 PM
>>>>>>>> To: "users@httpd.apache.org <mailto:users@httpd.apache.org>"
>>>>>>>> <users@httpd.apache.org <mailto:users@httpd.apache.org>>
>>>>>>>> Subject: Re: [users@httpd] Circumstances when mod_php would run
>>>>>>>> faster
>>>>>>>> than PHP-FPM?
>>>>>>>>
>>>>>>>> Mod_proxy_fcgi + php-fpm or mod_fcgid with php fcgi sapi should
>>>>>>>>both
>>>>>>>> be
>>>>>>>> equivalent when tuned correctly.
>>>>>>>>
>>>>>>>> Your only option for running php in process efficiently is to use
>>>>>>>> the
>>>>>>>> non-threadsafe php in the httpd preform module. Your only option
>>>>>>>> for
>>>>>>>> running httpd efficiently is the event, or at least the worker
>>>>>>>>mpm.
>>>>>>>>
>>>>>>>> Since usually only a subset of the http requests are to a php
>>>>>>>> resource,
>>>>>>>> the answer is almost always 'no'.
>>>>>>>>
>>>>>>>> On Jan 8, 2016 16:48, "Rose, John B" <jbr...@utk.edu
>>>>>>>> <mailto:jbr...@utk.edu>> wrote:
>>>>>>>>
>>>>>>>> Apache 2.4
>>>>>>>>
>>>>>>>> On the same system, same web site, are there circumstances
>>>>>>>>when
>>>>>>>> Apache using mod_php would run faster than Apache using
>>>>>>>>PHP-FPM?
>>>>>>>>
>>>>>>>> Thanks
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Rich Bowen - rbo...@rcbowen.com - @rbowen
>>>>>>> http://apachecon.com/ - @apachecon
>>>>>>>
>>>>>>>
>>>>>>>--------------------------------------------------------------------
>>>>>>>-
>>>>>>> To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org
>>>>>>> For additional commands, e-mail: users-h...@httpd.apache.org
>>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>---------------------------------------------------------------------
>>>>>> To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org
>>>>>> For additional commands, e-mail: users-h...@httpd.apache.org
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Rich Bowen - rbo...@rcbowen.com - @rbowen
>>>>> http://apachecon.com/ - @apachecon
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org
>>>>> For additional commands, e-mail: users-h...@httpd.apache.org
>>>>>
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org
>>>> For additional commands, e-mail: users-h...@httpd.apache.org
>>>>
>>>
>>>
>>> --
>>> Rich Bowen - rbo...@rcbowen.com - @rbowen
>>> http://apachecon.com/ - @apachecon
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org
>>> For additional commands, e-mail: users-h...@httpd.apache.org
>>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org
>> For additional commands, e-mail: users-h...@httpd.apache.org
>>
>
>
>--
>Rich Bowen - rbo...@rcbowen.com - @rbowen
>http://apachecon.com/ - @apachecon
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org
>For additional commands, e-mail: users-h...@httpd.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org
For additional commands, e-mail: users-h...@httpd.apache.org
