Welcome! Log In Create A New Profile

Advanced

php-fpm log errors....

Posted by iberkner 
php-fpm log errors....
August 20, 2009 01:39PM
Hi All,

Getting these errors in my php-fpm log on an on-going basis. Questions:

a. Is this normal, i.e. can I expect these in such frequency?

b. How do I determine the cause?

c. Maybe there is no error? i.e. the php process has nothing to do so it
just shuts down?

Thanks!


Aug 20 12:33:48.220864 [NOTICE] fpm_children_make(), line 352: child 11428
(pool www-data) started
Aug 20 12:34:15.043999 [NOTICE] fpm_got_signal(), line 48: received SIGCHLD
Aug 20 12:34:15.044076 [NOTICE] fpm_children_bury(), line 215: child 11405
(pool www-data) exited with code 0 after 162.301219 seconds from start
Aug 20 12:34:15.045189 [NOTICE] fpm_children_make(), line 352: child 11429
(pool www-data) started
Aug 20 12:34:16.628762 [NOTICE] fpm_got_signal(), line 48: received SIGCHLD
Aug 20 12:34:16.628835 [NOTICE] fpm_children_bury(), line 215: child 11406
(pool www-data) exited with code 0 after 157.862689 seconds from start
Aug 20 12:34:16.630008 [NOTICE] fpm_children_make(), line 352: child 11430
(pool www-data) started
Aug 20 12:34:27.768232 [NOTICE] fpm_got_signal(), line 48: received SIGCHLD
Aug 20 12:34:27.768314 [NOTICE] fpm_children_bury(), line 215: child 11410
(pool www-data) exited with code 0 after 161.161887 seconds from start
Aug 20 12:34:27.768829 [NOTICE] fpm_children_make(), line 352: child 11432
(pool www-data) started
Re: php-fpm log errors....
August 20, 2009 01:41PM
no error there. that's normal it's recycling the children. however,
every 160 seconds is pretty quick. you could bump up the number of
engines or how many max requests each one does. i'd probably bump both
up.


On Thu, Aug 20, 2009 at 10:39 AM, Ilan Berkner<iberkner@gmail.com> wrote:
> Hi All,
>
> Getting these errors in my php-fpm log on an on-going basis.  Questions:
>
> a. Is this normal, i.e. can I expect these in such frequency?
>
> b. How do I determine the cause?
>
> c. Maybe there is no error? i.e. the php process has nothing to do so it
> just shuts down?
>
> Thanks!
>
>
> Aug 20 12:33:48.220864 [NOTICE] fpm_children_make(), line 352: child 11428
> (pool www-data) started
> Aug 20 12:34:15.043999 [NOTICE] fpm_got_signal(), line 48: received SIGCHLD
> Aug 20 12:34:15.044076 [NOTICE] fpm_children_bury(), line 215: child 11405
> (pool www-data) exited with code 0 after 162.301219 seconds from start
> Aug 20 12:34:15.045189 [NOTICE] fpm_children_make(), line 352: child 11429
> (pool www-data) started
> Aug 20 12:34:16.628762 [NOTICE] fpm_got_signal(), line 48: received SIGCHLD
> Aug 20 12:34:16.628835 [NOTICE] fpm_children_bury(), line 215: child 11406
> (pool www-data) exited with code 0 after 157.862689 seconds from start
> Aug 20 12:34:16.630008 [NOTICE] fpm_children_make(), line 352: child 11430
> (pool www-data) started
> Aug 20 12:34:27.768232 [NOTICE] fpm_got_signal(), line 48: received SIGCHLD
> Aug 20 12:34:27.768314 [NOTICE] fpm_children_bury(), line 215: child 11410
> (pool www-data) exited with code 0 after 161.161887 seconds from start
> Aug 20 12:34:27.768829 [NOTICE] fpm_children_make(), line 352: child 11432
> (pool www-data) started
>
>
Re: php-fpm log errors....
August 20, 2009 03:17PM
I bumped up the number of requests from 500 to 1000 and the number of
engines to 6 from 5.

Do you think that 1,000 is to high? Is 6 to low? What criteria do you use
to determine best settings?

Thanks

On Thu, Aug 20, 2009 at 1:41 PM, Michael Shadle wrote:

>
> no error there. that's normal it's recycling the children. however,
> every 160 seconds is pretty quick. you could bump up the number of
> engines or how many max requests each one does. i'd probably bump both
> up.
>
>
> On Thu, Aug 20, 2009 at 10:39 AM, Ilan Berkner wrote:
> > Hi All,
> >
> > Getting these errors in my php-fpm log on an on-going basis. Questions:
> >
> > a. Is this normal, i.e. can I expect these in such frequency?
> >
> > b. How do I determine the cause?
> >
> > c. Maybe there is no error? i.e. the php process has nothing to do so it
> > just shuts down?
> >
> > Thanks!
> >
> >
> > Aug 20 12:33:48.220864 [NOTICE] fpm_children_make(), line 352: child
> 11428
> > (pool www-data) started
> > Aug 20 12:34:15.043999 [NOTICE] fpm_got_signal(), line 48: received
> SIGCHLD
> > Aug 20 12:34:15.044076 [NOTICE] fpm_children_bury(), line 215: child
> 11405
> > (pool www-data) exited with code 0 after 162.301219 seconds from start
> > Aug 20 12:34:15.045189 [NOTICE] fpm_children_make(), line 352: child
> 11429
> > (pool www-data) started
> > Aug 20 12:34:16.628762 [NOTICE] fpm_got_signal(), line 48: received
> SIGCHLD
> > Aug 20 12:34:16.628835 [NOTICE] fpm_children_bury(), line 215: child
> 11406
> > (pool www-data) exited with code 0 after 157.862689 seconds from start
> > Aug 20 12:34:16.630008 [NOTICE] fpm_children_make(), line 352: child
> 11430
> > (pool www-data) started
> > Aug 20 12:34:27.768232 [NOTICE] fpm_got_signal(), line 48: received
> SIGCHLD
> > Aug 20 12:34:27.768314 [NOTICE] fpm_children_bury(), line 215: child
> 11410
> > (pool www-data) exited with code 0 after 161.161887 seconds from start
> > Aug 20 12:34:27.768829 [NOTICE] fpm_children_make(), line 352: child
> 11432
> > (pool www-data) started
> >
> >
>
Re: php-fpm log errors....
August 20, 2009 03:29PM
i've used 125, 250, and 500. i think 500 is default typically. it just
depends on your resources.

basically figure out that engines * requests = how many requests your
pool can handle before it will recycle, i believe. i don't think it
will recycle individual children. or i'm completely confused right now
(i'm at a conference with lagged wifi and i can't concentrate enough
to examine things.)



On Thu, Aug 20, 2009 at 12:17 PM, Ilan Berkner<iberkner@gmail.com> wrote:
> I bumped up the number of requests from 500 to 1000 and the number of
> engines to 6 from 5.
>
> Do you think that 1,000 is to high?  Is 6 to low?  What criteria do you use
> to determine best settings?
>
> Thanks
>
> On Thu, Aug 20, 2009 at 1:41 PM, Michael Shadle <mike503@gmail.com> wrote:
>>
>> no error there. that's normal it's recycling the children. however,
>> every 160 seconds is pretty quick. you could bump up the number of
>> engines or how many max requests each one does. i'd probably bump both
>> up.
>>
>>
>> On Thu, Aug 20, 2009 at 10:39 AM, Ilan Berkner<iberkner@gmail.com> wrote:
>> > Hi All,
>> >
>> > Getting these errors in my php-fpm log on an on-going basis.  Questions:
>> >
>> > a. Is this normal, i.e. can I expect these in such frequency?
>> >
>> > b. How do I determine the cause?
>> >
>> > c. Maybe there is no error? i.e. the php process has nothing to do so it
>> > just shuts down?
>> >
>> > Thanks!
>> >
>> >
>> > Aug 20 12:33:48.220864 [NOTICE] fpm_children_make(), line 352: child
>> > 11428
>> > (pool www-data) started
>> > Aug 20 12:34:15.043999 [NOTICE] fpm_got_signal(), line 48: received
>> > SIGCHLD
>> > Aug 20 12:34:15.044076 [NOTICE] fpm_children_bury(), line 215: child
>> > 11405
>> > (pool www-data) exited with code 0 after 162.301219 seconds from start
>> > Aug 20 12:34:15.045189 [NOTICE] fpm_children_make(), line 352: child
>> > 11429
>> > (pool www-data) started
>> > Aug 20 12:34:16.628762 [NOTICE] fpm_got_signal(), line 48: received
>> > SIGCHLD
>> > Aug 20 12:34:16.628835 [NOTICE] fpm_children_bury(), line 215: child
>> > 11406
>> > (pool www-data) exited with code 0 after 157.862689 seconds from start
>> > Aug 20 12:34:16.630008 [NOTICE] fpm_children_make(), line 352: child
>> > 11430
>> > (pool www-data) started
>> > Aug 20 12:34:27.768232 [NOTICE] fpm_got_signal(), line 48: received
>> > SIGCHLD
>> > Aug 20 12:34:27.768314 [NOTICE] fpm_children_bury(), line 215: child
>> > 11410
>> > (pool www-data) exited with code 0 after 161.161887 seconds from start
>> > Aug 20 12:34:27.768829 [NOTICE] fpm_children_make(), line 352: child
>> > 11432
>> > (pool www-data) started
>> >
>> >
>
>
Re: php-fpm log errors....
August 20, 2009 03:32PM
thanks for the info ...

when you get a chance, if you can let me know how to figure that out
(engines * requests) that would be great.

Thanks

On Thu, Aug 20, 2009 at 3:29 PM, Michael Shadle wrote:

>
> i've used 125, 250, and 500. i think 500 is default typically. it just
> depends on your resources.
>
> basically figure out that engines * requests = how many requests your
> pool can handle before it will recycle, i believe. i don't think it
> will recycle individual children. or i'm completely confused right now
> (i'm at a conference with lagged wifi and i can't concentrate enough
> to examine things.)
>
>
>
> On Thu, Aug 20, 2009 at 12:17 PM, Ilan Berkner wrote:
> > I bumped up the number of requests from 500 to 1000 and the number of
> > engines to 6 from 5.
> >
> > Do you think that 1,000 is to high? Is 6 to low? What criteria do you
> use
> > to determine best settings?
> >
> > Thanks
> >
> > On Thu, Aug 20, 2009 at 1:41 PM, Michael Shadle
> wrote:
> >>
> >> no error there. that's normal it's recycling the children. however,
> >> every 160 seconds is pretty quick. you could bump up the number of
> >> engines or how many max requests each one does. i'd probably bump both
> >> up.
> >>
> >>
> >> On Thu, Aug 20, 2009 at 10:39 AM, Ilan Berkner
> wrote:
> >> > Hi All,
> >> >
> >> > Getting these errors in my php-fpm log on an on-going basis.
> Questions:
> >> >
> >> > a. Is this normal, i.e. can I expect these in such frequency?
> >> >
> >> > b. How do I determine the cause?
> >> >
> >> > c. Maybe there is no error? i.e. the php process has nothing to do so
> it
> >> > just shuts down?
> >> >
> >> > Thanks!
> >> >
> >> >
> >> > Aug 20 12:33:48.220864 [NOTICE] fpm_children_make(), line 352: child
> >> > 11428
> >> > (pool www-data) started
> >> > Aug 20 12:34:15.043999 [NOTICE] fpm_got_signal(), line 48: received
> >> > SIGCHLD
> >> > Aug 20 12:34:15.044076 [NOTICE] fpm_children_bury(), line 215: child
> >> > 11405
> >> > (pool www-data) exited with code 0 after 162.301219 seconds from start
> >> > Aug 20 12:34:15.045189 [NOTICE] fpm_children_make(), line 352: child
> >> > 11429
> >> > (pool www-data) started
> >> > Aug 20 12:34:16.628762 [NOTICE] fpm_got_signal(), line 48: received
> >> > SIGCHLD
> >> > Aug 20 12:34:16.628835 [NOTICE] fpm_children_bury(), line 215: child
> >> > 11406
> >> > (pool www-data) exited with code 0 after 157.862689 seconds from start
> >> > Aug 20 12:34:16.630008 [NOTICE] fpm_children_make(), line 352: child
> >> > 11430
> >> > (pool www-data) started
> >> > Aug 20 12:34:27.768232 [NOTICE] fpm_got_signal(), line 48: received
> >> > SIGCHLD
> >> > Aug 20 12:34:27.768314 [NOTICE] fpm_children_bury(), line 215: child
> >> > 11410
> >> > (pool www-data) exited with code 0 after 161.161887 seconds from start
> >> > Aug 20 12:34:27.768829 [NOTICE] fpm_children_make(), line 352: child
> >> > 11432
> >> > (pool www-data) started
> >> >
> >> >
> >
> >
>
Vasiliy G Tolstov
Re: php-fpm log errors....
August 20, 2009 02:50PM
On Thu, 2009-08-20 at 13:39 -0400, Ilan Berkner wrote:
> Hi All,
>
> Getting these errors in my php-fpm log on an on-going basis.
> Questions:
>
> a. Is this normal, i.e. can I expect these in such frequency?
this is normal - worker proccess exited with ok status (0) and restarted
please see config file for requests count, after it process restarted to
avoid memory leaks.



--
Vasiliy G Tolstov <v.tolstov@selfip.ru>
Selfip.Ru
Re: php-fpm log errors....
August 20, 2009 03:51PM
My only confusion is if each child terminates on it's own or the
entire pool waits before it recycles.

Sent from my iPhone

On Aug 20, 2009, at 12:32 PM, Ilan Berkner wrote:

> thanks for the info ...
>
> when you get a chance, if you can let me know how to figure that out
> (engines * requests) that would be great.
>
> Thanks
>
> On Thu, Aug 20, 2009 at 3:29 PM, Michael Shadle
> wrote:
>
> i've used 125, 250, and 500. i think 500 is default typically. it just
> depends on your resources.
>
> basically figure out that engines * requests = how many requests your
> pool can handle before it will recycle, i believe. i don't think it
> will recycle individual children. or i'm completely confused right now
> (i'm at a conference with lagged wifi and i can't concentrate enough
> to examine things.)
>
>
>
> On Thu, Aug 20, 2009 at 12:17 PM, Ilan Berkner
> wrote:
> > I bumped up the number of requests from 500 to 1000 and the number
> of
> > engines to 6 from 5.
> >
> > Do you think that 1,000 is to high? Is 6 to low? What criteria
> do you use
> > to determine best settings?
> >
> > Thanks
> >
> > On Thu, Aug 20, 2009 at 1:41 PM, Michael Shadle
> wrote:
> >>
> >> no error there. that's normal it's recycling the children. however,
> >> every 160 seconds is pretty quick. you could bump up the number of
> >> engines or how many max requests each one does. i'd probably bump
> both
> >> up.
> >>
> >>
> >> On Thu, Aug 20, 2009 at 10:39 AM, Ilan
> Berkner wrote:
> >> > Hi All,
> >> >
> >> > Getting these errors in my php-fpm log on an on-going basis.
> Questions:
> >> >
> >> > a. Is this normal, i.e. can I expect these in such frequency?
> >> >
> >> > b. How do I determine the cause?
> >> >
> >> > c. Maybe there is no error? i.e. the php process has nothing to
> do so it
> >> > just shuts down?
> >> >
> >> > Thanks!
> >> >
> >> >
> >> > Aug 20 12:33:48.220864 [NOTICE] fpm_children_make(), line 352:
> child
> >> > 11428
> >> > (pool www-data) started
> >> > Aug 20 12:34:15.043999 [NOTICE] fpm_got_signal(), line 48:
> received
> >> > SIGCHLD
> >> > Aug 20 12:34:15.044076 [NOTICE] fpm_children_bury(), line 215:
> child
> >> > 11405
> >> > (pool www-data) exited with code 0 after 162.301219 seconds
> from start
> >> > Aug 20 12:34:15.045189 [NOTICE] fpm_children_make(), line 352:
> child
> >> > 11429
> >> > (pool www-data) started
> >> > Aug 20 12:34:16.628762 [NOTICE] fpm_got_signal(), line 48:
> received
> >> > SIGCHLD
> >> > Aug 20 12:34:16.628835 [NOTICE] fpm_children_bury(), line 215:
> child
> >> > 11406
> >> > (pool www-data) exited with code 0 after 157.862689 seconds
> from start
> >> > Aug 20 12:34:16.630008 [NOTICE] fpm_children_make(), line 352:
> child
> >> > 11430
> >> > (pool www-data) started
> >> > Aug 20 12:34:27.768232 [NOTICE] fpm_got_signal(), line 48:
> received
> >> > SIGCHLD
> >> > Aug 20 12:34:27.768314 [NOTICE] fpm_children_bury(), line 215:
> child
> >> > 11410
> >> > (pool www-data) exited with code 0 after 161.161887 seconds
> from start
> >> > Aug 20 12:34:27.768829 [NOTICE] fpm_children_make(), line 352:
> child
> >> > 11432
> >> > (pool www-data) started
> >> >
> >> >
> >
> >
>
Jim Ohlstein
Re: php-fpm log errors....
August 20, 2009 04:01PM
I'm fairly certain that each child is terminated individually after it handles the defined number of requests. Since over the lifetime of a pool of children requests are handled in *roughly* the same numbers it is likey that all children will be recycling in a reasonably short window but it's not necessarily always the case.

Sent from my Verizon Wireless BlackBerry

-----Original Message-----
From: Michael Shadle

Date: Thu, 20 Aug 2009 12:51:16
To: highload-php-en@googlegroups.com
Cc: highload-php-en@googlegroups.com
Subject: Re: php-fpm log errors....


My only confusion is if each child terminates on it's own or the
entire pool waits before it recycles.

Sent from my iPhone

On Aug 20, 2009, at 12:32 PM, Ilan Berkner wrote:

> thanks for the info ...
>
> when you get a chance, if you can let me know how to figure that out
> (engines * requests) that would be great.
>
> Thanks
>
> On Thu, Aug 20, 2009 at 3:29 PM, Michael Shadle
> wrote:
>
> i've used 125, 250, and 500. i think 500 is default typically. it just
> depends on your resources.
>
> basically figure out that engines * requests = how many requests your
> pool can handle before it will recycle, i believe. i don't think it
> will recycle individual children. or i'm completely confused right now
> (i'm at a conference with lagged wifi and i can't concentrate enough
> to examine things.)
>
>
>
> On Thu, Aug 20, 2009 at 12:17 PM, Ilan Berkner
> wrote:
> > I bumped up the number of requests from 500 to 1000 and the number
> of
> > engines to 6 from 5.
> >
> > Do you think that 1,000 is to high? Is 6 to low? What criteria
> do you use
> > to determine best settings?
> >
> > Thanks
> >
> > On Thu, Aug 20, 2009 at 1:41 PM, Michael Shadle
> wrote:
> >>
> >> no error there. that's normal it's recycling the children. however,
> >> every 160 seconds is pretty quick. you could bump up the number of
> >> engines or how many max requests each one does. i'd probably bump
> both
> >> up.
> >>
> >>
> >> On Thu, Aug 20, 2009 at 10:39 AM, Ilan
> Berkner wrote:
> >> > Hi All,
> >> >
> >> > Getting these errors in my php-fpm log on an on-going basis.
> Questions:
> >> >
> >> > a. Is this normal, i.e. can I expect these in such frequency?
> >> >
> >> > b. How do I determine the cause?
> >> >
> >> > c. Maybe there is no error? i.e. the php process has nothing to
> do so it
> >> > just shuts down?
> >> >
> >> > Thanks!
> >> >
> >> >
> >> > Aug 20 12:33:48.220864 [NOTICE] fpm_children_make(), line 352:
> child
> >> > 11428
> >> > (pool www-data) started
> >> > Aug 20 12:34:15.043999 [NOTICE] fpm_got_signal(), line 48:
> received
> >> > SIGCHLD
> >> > Aug 20 12:34:15.044076 [NOTICE] fpm_children_bury(), line 215:
> child
> >> > 11405
> >> > (pool www-data) exited with code 0 after 162.301219 seconds
> from start
> >> > Aug 20 12:34:15.045189 [NOTICE] fpm_children_make(), line 352:
> child
> >> > 11429
> >> > (pool www-data) started
> >> > Aug 20 12:34:16.628762 [NOTICE] fpm_got_signal(), line 48:
> received
> >> > SIGCHLD
> >> > Aug 20 12:34:16.628835 [NOTICE] fpm_children_bury(), line 215:
> child
> >> > 11406
> >> > (pool www-data) exited with code 0 after 157.862689 seconds
> from start
> >> > Aug 20 12:34:16.630008 [NOTICE] fpm_children_make(), line 352:
> child
> >> > 11430
> >> > (pool www-data) started
> >> > Aug 20 12:34:27.768232 [NOTICE] fpm_got_signal(), line 48:
> received
> >> > SIGCHLD
> >> > Aug 20 12:34:27.768314 [NOTICE] fpm_children_bury(), line 215:
> child
> >> > 11410
> >> > (pool www-data) exited with code 0 after 161.161887 seconds
> from start
> >> > Aug 20 12:34:27.768829 [NOTICE] fpm_children_make(), line 352:
> child
> >> > 11432
> >> > (pool www-data) started
> >> >
> >> >
> >
> >
>
Re: php-fpm log errors....
August 20, 2009 04:28PM
I would put my money on that as well.

Also, it's not necessarily a bad thing I guess if it's recycling every
160 seconds. It does mean it is releasing memory a lot quicker from
possible PHP leaks quicker. However your APC cache is probably blown
away each time then too...

On Thu, Aug 20, 2009 at 1:01 PM, Jim Ohlstein<jim.ohlstein@gmail.com> wrote:
> I'm fairly certain that each child is terminated individually after it
> handles the defined number of requests. Since over the lifetime of a pool of
> children requests are handled in *roughly* the same numbers it is likey that
> all children will be recycling in a reasonably short window but it's not
> necessarily always the case.
>
> Sent from my Verizon Wireless BlackBerry
>
> ________________________________
> From: Michael Shadle
> Date: Thu, 20 Aug 2009 12:51:16 -0700
> To: highload-php-en@googlegroups.com<highload-php-en@googlegroups.com>
> Subject: Re: php-fpm log errors....
>
> My only confusion is if each child terminates on it's own or the entire pool
> waits before it recycles.
>
> Sent from my iPhone
> On Aug 20, 2009, at 12:32 PM, Ilan Berkner <iberkner@gmail.com> wrote:
>
> thanks for the info ...
>
> when you get a chance, if you can let me know how to figure that out
> (engines * requests) that would be great.
>
> Thanks
>
> On Thu, Aug 20, 2009 at 3:29 PM, Michael Shadle <mike503@gmail.com> wrote:
>>
>> i've used 125, 250, and 500. i think 500 is default typically. it just
>> depends on your resources.
>>
>> basically figure out that engines * requests = how many requests your
>> pool can handle before it will recycle, i believe. i don't think it
>> will recycle individual children. or i'm completely confused right now
>> (i'm at a conference with lagged wifi and i can't concentrate enough
>> to examine things.)
>>
>>
>>
>> On Thu, Aug 20, 2009 at 12:17 PM, Ilan Berkner<iberkner@gmail.com> wrote:
>> > I bumped up the number of requests from 500 to 1000 and the number of
>> > engines to 6 from 5.
>> >
>> > Do you think that 1,000 is to high?  Is 6 to low?  What criteria do you
>> > use
>> > to determine best settings?
>> >
>> > Thanks
>> >
>> > On Thu, Aug 20, 2009 at 1:41 PM, Michael Shadle <mike503@gmail.com>
>> > wrote:
>> >>
>> >> no error there. that's normal it's recycling the children. however,
>> >> every 160 seconds is pretty quick. you could bump up the number of
>> >> engines or how many max requests each one does. i'd probably bump both
>> >> up.
>> >>
>> >>
>> >> On Thu, Aug 20, 2009 at 10:39 AM, Ilan Berkner<iberkner@gmail.com>
>> >> wrote:
>> >> > Hi All,
>> >> >
>> >> > Getting these errors in my php-fpm log on an on-going basis.
>> >> > Questions:
>> >> >
>> >> > a. Is this normal, i.e. can I expect these in such frequency?
>> >> >
>> >> > b. How do I determine the cause?
>> >> >
>> >> > c. Maybe there is no error? i.e. the php process has nothing to do so
>> >> > it
>> >> > just shuts down?
>> >> >
>> >> > Thanks!
>> >> >
>> >> >
>> >> > Aug 20 12:33:48.220864 [NOTICE] fpm_children_make(), line 352: child
>> >> > 11428
>> >> > (pool www-data) started
>> >> > Aug 20 12:34:15.043999 [NOTICE] fpm_got_signal(), line 48: received
>> >> > SIGCHLD
>> >> > Aug 20 12:34:15.044076 [NOTICE] fpm_children_bury(), line 215: child
>> >> > 11405
>> >> > (pool www-data) exited with code 0 after 162.301219 seconds from
>> >> > start
>> >> > Aug 20 12:34:15.045189 [NOTICE] fpm_children_make(), line 352: child
>> >> > 11429
>> >> > (pool www-data) started
>> >> > Aug 20 12:34:16.628762 [NOTICE] fpm_got_signal(), line 48: received
>> >> > SIGCHLD
>> >> > Aug 20 12:34:16.628835 [NOTICE] fpm_children_bury(), line 215: child
>> >> > 11406
>> >> > (pool www-data) exited with code 0 after 157.862689 seconds from
>> >> > start
>> >> > Aug 20 12:34:16.630008 [NOTICE] fpm_children_make(), line 352: child
>> >> > 11430
>> >> > (pool www-data) started
>> >> > Aug 20 12:34:27.768232 [NOTICE] fpm_got_signal(), line 48: received
>> >> > SIGCHLD
>> >> > Aug 20 12:34:27.768314 [NOTICE] fpm_children_bury(), line 215: child
>> >> > 11410
>> >> > (pool www-data) exited with code 0 after 161.161887 seconds from
>> >> > start
>> >> > Aug 20 12:34:27.768829 [NOTICE] fpm_children_make(), line 352: child
>> >> > 11432
>> >> > (pool www-data) started
>> >> >
>> >> >
>> >
>> >
>
>
Jim Ohlstein
Re: php-fpm log errors....
August 20, 2009 04:44PM
Michael Shadle wrote:
> I would put my money on that as well.
>
> Also, it's not necessarily a bad thing I guess if it's recycling every
> 160 seconds. It does mean it is releasing memory a lot quicker from
> possible PHP leaks quicker. However your APC cache is probably blown
> away each time then too...
>
I don't know if that's the case since the master process is not being
recycled ever and they should all be sharing the same memory. Maybe I am
wrong? I use XCache in general and I routinely have over 1,000,000,000
cache "hits" between restarts. During that time php-cgi child processes
are recycled hundreds of thousands of times each.
> On Thu, Aug 20, 2009 at 1:01 PM, Jim Ohlstein<jim.ohlstein@gmail.com> wrote:
>
>> I'm fairly certain that each child is terminated individually after it
>> handles the defined number of requests. Since over the lifetime of a pool of
>> children requests are handled in *roughly* the same numbers it is likey that
>> all children will be recycling in a reasonably short window but it's not
>> necessarily always the case.
>>
>> Sent from my Verizon Wireless BlackBerry
>>
>> ________________________________
>> From: Michael Shadle
>> Date: Thu, 20 Aug 2009 12:51:16 -0700
>> To: highload-php-en@googlegroups.com<highload-php-en@googlegroups.com>
>> Subject: Re: php-fpm log errors....
>>
>> My only confusion is if each child terminates on it's own or the entire pool
>> waits before it recycles.
>>
>> Sent from my iPhone
>> On Aug 20, 2009, at 12:32 PM, Ilan Berkner <iberkner@gmail.com> wrote:
>>
>> thanks for the info ...
>>
>> when you get a chance, if you can let me know how to figure that out
>> (engines * requests) that would be great.
>>
>> Thanks
>>
>> On Thu, Aug 20, 2009 at 3:29 PM, Michael Shadle <mike503@gmail.com> wrote:
>>
>>> i've used 125, 250, and 500. i think 500 is default typically. it just
>>> depends on your resources.
>>>
>>> basically figure out that engines * requests = how many requests your
>>> pool can handle before it will recycle, i believe. i don't think it
>>> will recycle individual children. or i'm completely confused right now
>>> (i'm at a conference with lagged wifi and i can't concentrate enough
>>> to examine things.)
>>>
>>>
>>>
>>> On Thu, Aug 20, 2009 at 12:17 PM, Ilan Berkner<iberkner@gmail.com> wrote:
>>>
>>>> I bumped up the number of requests from 500 to 1000 and the number of
>>>> engines to 6 from 5.
>>>>
>>>> Do you think that 1,000 is to high? Is 6 to low? What criteria do you
>>>> use
>>>> to determine best settings?
>>>>
>>>> Thanks
>>>>
>>>> On Thu, Aug 20, 2009 at 1:41 PM, Michael Shadle <mike503@gmail.com>
>>>> wrote:
>>>>
>>>>> no error there. that's normal it's recycling the children. however,
>>>>> every 160 seconds is pretty quick. you could bump up the number of
>>>>> engines or how many max requests each one does. i'd probably bump both
>>>>> up.
>>>>>
>>>>>
>>>>> On Thu, Aug 20, 2009 at 10:39 AM, Ilan Berkner<iberkner@gmail.com>
>>>>> wrote:
>>>>>
>>>>>> Hi All,
>>>>>>
>>>>>> Getting these errors in my php-fpm log on an on-going basis.
>>>>>> Questions:
>>>>>>
>>>>>> a. Is this normal, i.e. can I expect these in such frequency?
>>>>>>
>>>>>> b. How do I determine the cause?
>>>>>>
>>>>>> c. Maybe there is no error? i.e. the php process has nothing to do so
>>>>>> it
>>>>>> just shuts down?
>>>>>>
>>>>>> Thanks!
>>>>>>
>>>>>>
>>>>>> Aug 20 12:33:48.220864 [NOTICE] fpm_children_make(), line 352: child
>>>>>> 11428
>>>>>> (pool www-data) started
>>>>>> Aug 20 12:34:15.043999 [NOTICE] fpm_got_signal(), line 48: received
>>>>>> SIGCHLD
>>>>>> Aug 20 12:34:15.044076 [NOTICE] fpm_children_bury(), line 215: child
>>>>>> 11405
>>>>>> (pool www-data) exited with code 0 after 162.301219 seconds from
>>>>>> start
>>>>>> Aug 20 12:34:15.045189 [NOTICE] fpm_children_make(), line 352: child
>>>>>> 11429
>>>>>> (pool www-data) started
>>>>>> Aug 20 12:34:16.628762 [NOTICE] fpm_got_signal(), line 48: received
>>>>>> SIGCHLD
>>>>>> Aug 20 12:34:16.628835 [NOTICE] fpm_children_bury(), line 215: child
>>>>>> 11406
>>>>>> (pool www-data) exited with code 0 after 157.862689 seconds from
>>>>>> start
>>>>>> Aug 20 12:34:16.630008 [NOTICE] fpm_children_make(), line 352: child
>>>>>> 11430
>>>>>> (pool www-data) started
>>>>>> Aug 20 12:34:27.768232 [NOTICE] fpm_got_signal(), line 48: received
>>>>>> SIGCHLD
>>>>>> Aug 20 12:34:27.768314 [NOTICE] fpm_children_bury(), line 215: child
>>>>>> 11410
>>>>>> (pool www-data) exited with code 0 after 161.161887 seconds from
>>>>>> start
>>>>>> Aug 20 12:34:27.768829 [NOTICE] fpm_children_make(), line 352: child
>>>>>> 11432
>>>>>> (pool www-data) started
>>>>>>
>>>>>>
>>>>>>
>>>>
>>
>
>
Re: php-fpm log errors....
August 20, 2009 04:59PM
yeah i am not sure. maybe it is all in the master. makes sense
otherwise children would be isolated.

On Thu, Aug 20, 2009 at 1:44 PM, Jim Ohlstein<jim.ohlstein@gmail.com> wrote:
>
>
>
> Michael Shadle wrote:
>>
>> I would put my money on that as well.
>>
>> Also, it's not necessarily a bad thing I guess if it's recycling every
>> 160 seconds. It does mean it is releasing memory a lot quicker from
>> possible PHP leaks quicker. However your APC cache is probably blown
>> away each time then too...
>>
>
> I don't know if that's the case since the master process is not being
> recycled ever and they should all be sharing the same memory. Maybe I am
> wrong? I use XCache in general and I routinely have over 1,000,000,000 cache
> "hits" between restarts. During that time php-cgi child processes are
> recycled hundreds of thousands of times each.
>>
>> On Thu, Aug 20, 2009 at 1:01 PM, Jim Ohlstein<jim.ohlstein@gmail.com>
>> wrote:
>>
>>>
>>> I'm fairly certain that each child is terminated individually after it
>>> handles the defined number of requests. Since over the lifetime of a pool
>>> of
>>> children requests are handled in *roughly* the same numbers it is likey
>>> that
>>> all children will be recycling in a reasonably short window but it's not
>>> necessarily always the case.
>>>
>>> Sent from my Verizon Wireless BlackBerry
>>>
>>> ________________________________
>>> From: Michael Shadle
>>> Date: Thu, 20 Aug 2009 12:51:16 -0700
>>> To: highload-php-en@googlegroups.com<highload-php-en@googlegroups.com>
>>> Subject: Re: php-fpm log errors....
>>>
>>> My only confusion is if each child terminates on it's own or the entire
>>> pool
>>> waits before it recycles.
>>>
>>> Sent from my iPhone
>>> On Aug 20, 2009, at 12:32 PM, Ilan Berkner <iberkner@gmail.com> wrote:
>>>
>>> thanks for the info ...
>>>
>>> when you get a chance, if you can let me know how to figure that out
>>> (engines * requests) that would be great.
>>>
>>> Thanks
>>>
>>> On Thu, Aug 20, 2009 at 3:29 PM, Michael Shadle <mike503@gmail.com>
>>> wrote:
>>>
>>>>
>>>> i've used 125, 250, and 500. i think 500 is default typically. it just
>>>> depends on your resources.
>>>>
>>>> basically figure out that engines * requests = how many requests your
>>>> pool can handle before it will recycle, i believe. i don't think it
>>>> will recycle individual children. or i'm completely confused right now
>>>> (i'm at a conference with lagged wifi and i can't concentrate enough
>>>> to examine things.)
>>>>
>>>>
>>>>
>>>> On Thu, Aug 20, 2009 at 12:17 PM, Ilan Berkner<iberkner@gmail.com>
>>>> wrote:
>>>>
>>>>>
>>>>> I bumped up the number of requests from 500 to 1000 and the number of
>>>>> engines to 6 from 5.
>>>>>
>>>>> Do you think that 1,000 is to high?  Is 6 to low?  What criteria do you
>>>>> use
>>>>> to determine best settings?
>>>>>
>>>>> Thanks
>>>>>
>>>>> On Thu, Aug 20, 2009 at 1:41 PM, Michael Shadle <mike503@gmail.com>
>>>>> wrote:
>>>>>
>>>>>>
>>>>>> no error there. that's normal it's recycling the children. however,
>>>>>> every 160 seconds is pretty quick. you could bump up the number of
>>>>>> engines or how many max requests each one does. i'd probably bump both
>>>>>> up.
>>>>>>
>>>>>>
>>>>>> On Thu, Aug 20, 2009 at 10:39 AM, Ilan Berkner<iberkner@gmail.com>
>>>>>> wrote:
>>>>>>
>>>>>>>
>>>>>>> Hi All,
>>>>>>>
>>>>>>> Getting these errors in my php-fpm log on an on-going basis.
>>>>>>> Questions:
>>>>>>>
>>>>>>> a. Is this normal, i.e. can I expect these in such frequency?
>>>>>>>
>>>>>>> b. How do I determine the cause?
>>>>>>>
>>>>>>> c. Maybe there is no error? i.e. the php process has nothing to do so
>>>>>>> it
>>>>>>> just shuts down?
>>>>>>>
>>>>>>> Thanks!
>>>>>>>
>>>>>>>
>>>>>>> Aug 20 12:33:48.220864 [NOTICE] fpm_children_make(), line 352: child
>>>>>>> 11428
>>>>>>> (pool www-data) started
>>>>>>> Aug 20 12:34:15.043999 [NOTICE] fpm_got_signal(), line 48: received
>>>>>>> SIGCHLD
>>>>>>> Aug 20 12:34:15.044076 [NOTICE] fpm_children_bury(), line 215: child
>>>>>>> 11405
>>>>>>> (pool www-data) exited with code 0 after 162.301219 seconds from
>>>>>>> start
>>>>>>> Aug 20 12:34:15.045189 [NOTICE] fpm_children_make(), line 352: child
>>>>>>> 11429
>>>>>>> (pool www-data) started
>>>>>>> Aug 20 12:34:16.628762 [NOTICE] fpm_got_signal(), line 48: received
>>>>>>> SIGCHLD
>>>>>>> Aug 20 12:34:16.628835 [NOTICE] fpm_children_bury(), line 215: child
>>>>>>> 11406
>>>>>>> (pool www-data) exited with code 0 after 157.862689 seconds from
>>>>>>> start
>>>>>>> Aug 20 12:34:16.630008 [NOTICE] fpm_children_make(), line 352: child
>>>>>>> 11430
>>>>>>> (pool www-data) started
>>>>>>> Aug 20 12:34:27.768232 [NOTICE] fpm_got_signal(), line 48: received
>>>>>>> SIGCHLD
>>>>>>> Aug 20 12:34:27.768314 [NOTICE] fpm_children_bury(), line 215: child
>>>>>>> 11410
>>>>>>> (pool www-data) exited with code 0 after 161.161887 seconds from
>>>>>>> start
>>>>>>> Aug 20 12:34:27.768829 [NOTICE] fpm_children_make(), line 352: child
>>>>>>> 11432
>>>>>>> (pool www-data) started
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>
>>>>>
>>>
>>>
>>
>>
>
Sorry, only registered users may post in this forum.

Click here to login

Online Users

Guests: 121
Record Number of Users: 8 on April 13, 2023
Record Number of Guests: 500 on July 15, 2024
Powered by nginx      Powered by FreeBSD      PHP Powered      Powered by MariaDB      ipv6 ready