[peruser] Patch for busy servers
janno at kood.ee
Tue Jan 4 08:30:48 MST 2011
Graceful seems to be really broken.
after 4 days of use we have ~100 000 lines in error_log stating:
[Sat Jan 01 23:12:52 2011] [warn] long lost child came home! (pid 32616).
This happens right after graceful restart.
The new peruser was installed 31 dec. and graceful restart was done few
hours later. 4 days later there were no new graceful restarts or regular
Yesterday the apache hung and wasn't responding anymore.
We have ~150 vhosts.
On 4.01.2011 17:17, Andrew wrote:
> Many thanks for your patches, they are appreciated. We will be testing this patch in the coming months.
> On 4 Jan 2011, at 15:12, Marcelo Coelho wrote:
>> I've made a few changes, so here is another update (v4). Probably the last, for while. :-)
>> Changes (from RC2):
>> * (v4) Code cleanup
>> * (v4) Performance: childs are started in ~25ms, 40 times faster than in RC2 (~1000ms)
>> * (v4) Bug fixed: now checking if total_processors is 1 (first access) to start StartProcessors
>> * (v3) Performance: new child type (CHILD_TYPE_RESERVED) to avoid collision (2 childs trying to get the same free slot)
>> * (v3) Bug fixed: fixed a bug in RC2, wait_timeout was always 0, never sleeping to wait for new workers.
>> * (v2) Performance: StartProcessors, new configuration directive to control the number of child processors per vhost at startup
>> * (v2) Performance: childs are started in ~50ms, 20 times faster than in RC2 (~1000ms)
>> * (v1) Performance: faster to lookup for free slots (this is important on busy servers, with many virtual hosts)
>> * (v1) Performance: faster to count processors, one single loop counts all processors
>> * (v1) Bug fixed: bug when MinSpareProcessors is set to 0 (now all workers processes are killed when idle_timeout is reached)
>> * (v1) Bug fixed: Free-up slots when a WORKER or PROCESSOR unexpectedly dies
>> Patch from RC2:
>> Full patch from vanilla Apache 2.2.17:
>> Marcelo Coelho
>> marcelo at mco2.com.br
>> Peruser mailing list
>> Peruser at telana.com
> Peruser mailing list
> Peruser at telana.com
More information about the Peruser