httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeff Trawick <traw...@gmail.com>
Subject Re: [users@httpd] Windows Apache 2.4.9 restarts itself
Date Tue, 05 Aug 2014 14:37:10 GMT
On Tue, Aug 5, 2014 at 8:25 AM, Jeff Trawick <trawick@gmail.com> wrote:

> On Tue, Aug 5, 2014 at 6:35 AM, Mar Imp <marimpimp@gmail.com> wrote:
>
>> Hi Jeff,
>>
>> after 5 Days the next Apache restart
>>
>> it contains a new Error:
>> [Tue Aug 05 01:36:30.138535 2014] [mpm_winnt:error] [pid 31884:tid
>> 444] (OS 6)The handle is invalid.  : AH00360: Child: Failure releasing
>> the start mutex
>>
>> Logfile:
>> [Tue Aug 05 01:36:29.137575 2014] [mpm_winnt:crit] [pid 31884:tid 444]
>> (OS 6)The handle is invalid.  : AH00356: Child: WAIT_FAILED --
>> shutting down server
>> [Tue Aug 05 01:36:29.137575 2014] [mpm_winnt:crit] [pid 31884:tid 444]
>> (OS 6)The handle is invalid.  : AH02644: Child: Event handle #0 (8) is
>> invalid
>> [Tue Aug 05 01:36:29.925336 2014] [mpm_winnt:warn] [pid 31884:tid 536]
>> (OS 10038)An operation was attempted on something that is not a
>> socket.  : AH00344: accept() failed.
>> [Tue Aug 05 01:36:30.138535 2014] [mpm_winnt:error] [pid 31884:tid
>> 444] (OS 6)The handle is invalid.  : AH00360: Child: Failure releasing
>> the start mutex
>> [Tue Aug 05 01:36:45.466468 2014] [mpm_winnt:warn] [pid 31884:tid
>> 5696] (OS 10038)An operation was attempted on something that is not a
>> socket.  : AH00344: accept() failed.
>> [Tue Aug 05 01:37:01.077535 2014] [mpm_winnt:notice] [pid 31884:tid
>> 444] AH00362: Child: Waiting 30 more seconds for 9 worker threads to
>> finish.
>> [Tue Aug 05 01:37:31.113122 2014] [mpm_winnt:notice] [pid 31884:tid
>> 444] AH00362: Child: Waiting 0 more seconds for 6 worker threads to
>> finish.
>> [Tue Aug 05 01:37:31.213217 2014] [mpm_winnt:notice] [pid 31884:tid
>> 444] AH00363: Child: Terminating 6 threads that failed to exit.
>> [Tue Aug 05 01:37:31.213217 2014] [mpm_winnt:notice] [pid 31884:tid
>> 444] AH00364: Child: All worker threads have exited.
>> [Tue Aug 05 01:37:31.370367 2014] [mpm_winnt:notice] [pid 13424:tid
>> 536] AH00428: Parent: child process 31884 exited with status 0 --
>> Restarting.
>> [Tue Aug 05 01:37:32.213174 2014] [mpm_winnt:notice] [pid 13424:tid
>> 536] AH00455: Apache/2.4.10 (Win64) OpenSSL/1.0.1h configured --
>> resuming normal operations
>> [Tue Aug 05 01:37:32.213174 2014] [mpm_winnt:notice] [pid 13424:tid
>> 536] AH00456: Apache Lounge VC11 Server built: Jul 17 2014 12:58:29
>> [Tue Aug 05 01:37:32.213174 2014] [core:notice] [pid 13424:tid 536]
>> AH00094: Command line: 'C:\\Apache24\\bin\\httpd.exe -d C:/Apache24'
>> [Tue Aug 05 01:37:32.214178 2014] [mpm_winnt:notice] [pid 13424:tid
>> 536] AH00418: Parent: Created child process 31668
>> [Tue Aug 05 01:37:33.568475 2014] [mpm_winnt:notice] [pid 31668:tid
>> 396] AH00354: Child: Starting 300 worker threads.
>>
>> Did you have time to investigate any further?
>> Can I help in any way ?
>>
>
> I am spending a modest amount of time on this.  For now, keep posting when
> you see something different in the log at the time of failure.
>
>
>
>>
>> thx and kind regards
>> Mar Imp
>>
>>
>> 2014-07-31 15:44 GMT+02:00 Mar Imp <marimpimp@gmail.com>:
>> > Hi,
>> >
>> > -> the Apache server restarted itself again:
>> >
>> > [Wed Jul 30 22:15:21.604987 2014] [mpm_winnt:crit] [pid 34808:tid 464]
>> > (OS 6)The handle is invalid.  : AH00356: Child: WAIT_FAILED --
>> > shutting down server
>> > [Wed Jul 30 22:15:21.604987 2014] [mpm_winnt:crit] [pid 34808:tid 464]
>> > (OS 6)The handle is invalid.  : AH02644: Child: Event handle #0 (8) is
>> > invalid
>> > [Wed Jul 30 22:15:23.026349 2014] [mpm_winnt:warn] [pid 34808:tid
>> > 4624] (OS 10038)An operation was attempted on something that is not a
>> > socket.  : AH00344: accept() failed.
>> > [Wed Jul 30 22:15:39.956951 2014] [mpm_winnt:warn] [pid 34808:tid
>> > 3092] (OS 10038)An operation was attempted on something that is not a
>> > socket.  : AH00344: accept() failed.
>> > [Wed Jul 30 22:15:53.544739 2014] [mpm_winnt:notice] [pid 34808:tid
>> > 464] AH00362: Child: Waiting 30 more seconds for 3 worker threads to
>> > finish.
>> > [Wed Jul 30 22:16:16.571534 2014] [mpm_winnt:notice] [pid 34808:tid
>> > 464] AH00364: Child: All worker threads have exited.
>> > [Wed Jul 30 22:16:16.687244 2014] [mpm_winnt:notice] [pid 13424:tid
>> > 536] AH00428: Parent: child process 34808 exited with status 0 --
>> > Restarting.
>> > [Wed Jul 30 22:16:17.454981 2014] [mpm_winnt:notice] [pid 13424:tid
>> > 536] AH00455: Apache/2.4.10 (Win64) OpenSSL/1.0.1h configured --
>> > resuming normal operations
>> > [Wed Jul 30 22:16:17.454981 2014] [mpm_winnt:notice] [pid 13424:tid
>> > 536] AH00456: Apache Lounge VC11 Server built: Jul 17 2014 12:58:29
>> > [Wed Jul 30 22:16:17.454981 2014] [core:notice] [pid 13424:tid 536]
>> > AH00094: Command line: 'C:\\Apache24\\bin\\httpd.exe -d C:/Apache24'
>> > [Wed Jul 30 22:16:17.455982 2014] [mpm_winnt:notice] [pid 13424:tid
>> > 536] AH00418: Parent: Created child process 31884
>> > [Wed Jul 30 22:16:18.727198 2014] [mpm_winnt:notice] [pid 31884:tid
>> > 444] AH00354: Child: Starting 300 worker threads.
>> >
>> > -> There are no suspicious entries in other apache logfiles or in the
>> > windows event log.
>> >
>> > -> you asked about loaded modules:
>> >
>> > C:\Apache24\bin>httpd.exe -t -D DUMP_MODULES
>> > Loaded Modules:
>> >  core_module (static)
>> >  win32_module (static)
>> >  mpm_winnt_module (static)
>> >  http_module (static)
>> >  so_module (static)
>> >  access_compat_module (shared)
>> >  actions_module (shared)
>> >  alias_module (shared)
>> >  allowmethods_module (shared)
>> >  asis_module (shared)
>> >  auth_basic_module (shared)
>> >  authn_core_module (shared)
>> >  authn_file_module (shared)
>> >  authz_core_module (shared)
>> >  authz_groupfile_module (shared)
>> >  authz_host_module (shared)
>> >  authz_user_module (shared)
>> >  autoindex_module (shared)
>> >  cache_module (shared)
>> >  cache_disk_module (shared)
>> >  cgi_module (shared)
>> >  deflate_module (shared)
>> >  dir_module (shared)
>> >  env_module (shared)
>> >  expires_module (shared)
>> >  filter_module (shared)
>> >  headers_module (shared)
>> >  include_module (shared)
>> >  isapi_module (shared)
>> >  log_config_module (shared)
>> >  mime_module (shared)
>> >  negotiation_module (shared)
>> >  proxy_module (shared)
>> >  proxy_connect_module (shared)
>> >  proxy_http_module (shared)
>> >  rewrite_module (shared)
>> >  setenvif_module (shared)
>> >  socache_shmcb_module (shared)
>> >  ssl_module (shared)
>> >
>> > -> apache works as a caching reverse proxy
>> >
>> > hope you can find out something that solves this issue ...
>> >
>> > thank you for your help!
>> >
>> > 2014-07-31 13:02 GMT+02:00 Jeff Trawick <trawick@gmail.com>:
>> >> On Tue, Jul 29, 2014 at 7:40 AM, Jeff Trawick <trawick@gmail.com>
>> wrote:
>> >>>
>> >>> On Tue, Jul 29, 2014 at 5:33 AM, Mar Imp <marimpimp@gmail.com>
wrote:
>> >>>>
>> >>>> Hi Jeff,
>> >>>>
>> >>>> After the first restart, apache restarted itself 2 more times
>> >>>> yesterday. Good news: the additional information IS part of the
logs:
>> >>>> [Mon Jul 28 16:48:10.070355 2014] [mpm_winnt:crit] [pid 38752:tid
>> 388]
>> >>>> (OS 6)The handle is invalid.  : AH02644: Child: Event handle #0
(8)
>> is
>> >>>> invalid
>> >>>> does it help you to identify the issue?
>> >>>
>> >>>
>> >>> Unfortunately it does not immediately pinpoint the problem, but it
>> can't
>> >>> hurt :)  8 is the value for the handle that suddenly went bad.
>> >>
>> >>
>> >> I'll try to spend some real time on this over the weekend.  This is a
>> very
>> >> interesting issue.
>> >>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
>> For additional commands, e-mail: users-help@httpd.apache.org
>>
>>
>
Here's the little bit I know:

* Value 8 for event handle #0 (exit_event) is fine.  (That's what I see in
my testing right after initialization.)
* Ignore AH00344 errors, which happen after shutdown starts.  These should
have a lower log priority if the server is already shutting down.
* If the exit event is closed inadvertently (tested by me with a patch to
close the event after a certain number of requests), the  shutdown starts
with the same error.
* I managed to wedge the server after a certain number of automatic
restarts of the child.

It would be helpful to check if some operational error or other relatively
infrequent activity is logged while handling a request right before the
child starts shutting down.  If you have separate error logs for vhosts,
check in those logs.

Mime
View raw message