Sorry. It isn't the socket which is the issue, but the lock. Locks don't 
generally work on NFS.

Same question though, is /web/logs really on an NFS partition?

You can try setting to a different lock mechanism. The choices you have 
available depend on the platform. See choices in:

  http://httpd.apache.org/docs/2.2/mod/mpm_common.html#acceptmutex

Rather than set AcceptMutex though, set WSGIAcceptMutex directive.

For example:

    WSGIAcceptMutex posixsem

I have not idea what BSD may support.

On 04/09/2014, at 5:16 PM, Graham Dumpleton <[email protected]> wrote:

> Is /web/logs really on a NFS partition.
> 
> I recollect that UNIX sockets may not work if located on a NFS partition.
> 
> Try using:
> 
>   WSGISocketPrefix /tmp/wsgi
> 
> Do note I don't recommend it being in /tmp, but is a good enough place that 
> should be on local partition to try with.
> 
> Graham
> 
> On 04/09/2014, at 4:33 PM, Tim Arnold <[email protected]> wrote:
> 
>> Hi Graham,
>> Thanks for your quick reply. 
>> I set home=/web/logs just to give a location because I always got a warning 
>> when I started httpd; 
>> I found that solution somewhere on the net (I think it was an answer you 
>> gave, but might be wrong).
>> 
>> The directory permission surprised me with the group setting, but here it is:
>> ls -las
>> total 156
>>   4 drwxrwxrwx  2 www   stats   4096 Sep  4 19:27 ./
>>   4 drwxrwxrwt  8 root  wheel   4096 Nov  1  2012 ../
>>   0 -rw-r--r--  1 root  wheel      0 Sep  4 19:27 .nfs.00a849a1.079b4.4
>>   0 -rw-------  1 root  wheel      0 Sep  4 19:27 accept.lock.1947
>>  48 -rw-r--r--  1 root  wheel  42942 Sep  4 18:58 access_log
>> 100 -rw-r--r--  1 root  wheel  97319 Sep  4 19:27 error_log
>>   0 -rw-r--r--  1 root  wheel      5 Sep  4 19:27 httpd.pid
>>   0 srwx------  1 www   wheel      0 Sep  4 19:27 wsgi.1947.0.1.sock=
>> 
>> This is a FreeBSD machine (amd64).
>> httpd -V
>> Server version: Apache/2.2.17 (Unix)
>> Server built:   Apr  5 2011 16:04:11
>> Server's Module Magic Number: 20051115:25
>> Server loaded:  APR 1.4.2, APR-Util 1.3.10
>> Compiled using: APR 1.4.2, APR-Util 1.3.10
>> Architecture:   64-bit
>> Server MPM:     Prefork
>>   threaded:     no
>>     forked:     yes (variable process count)
>> Server compiled with....
>>  -D APACHE_MPM_DIR="server/mpm/prefork"
>>  -D APR_HAS_SENDFILE
>>  -D APR_HAS_MMAP
>>  -D APR_HAVE_IPV6 (IPv4-mapped addresses enabled)
>>  -D APR_USE_FLOCK_SERIALIZE
>>  -D APR_USE_PTHREAD_SERIALIZE
>>  -D SINGLE_LISTEN_UNSERIALIZED_ACCEPT
>>  -D APR_HAS_OTHER_CHILD
>>  -D AP_HAVE_RELIABLE_PIPED_LOGS
>>  -D DYNAMIC_MODULE_LIMIT=128
>>  -D HTTPD_ROOT="/AppDocs/local/"
>>  -D SUEXEC_BIN="/AppDocs/local//bin/suexec"
>>  -D DEFAULT_PIDLOG="logs/httpd.pid"
>>  -D DEFAULT_SCOREBOARD="logs/apache_runtime_status"
>>  -D DEFAULT_LOCKFILE="logs/accept.lock"
>>  -D DEFAULT_ERRORLOG="logs/error_log"
>>  -D AP_TYPES_CONFIG_FILE="conf/mime.types"
>>  -D SERVER_CONFIG_FILE="conf/httpd.conf"
>> thanks again. The error messages start as soon as I start httpd (no pages 
>> requested).
>> --Tim
>> 
>> 
>> On Thursday, September 4, 2014 6:17:14 PM UTC-4, Graham Dumpleton wrote:
>> Why are you setting:
>> 
>>   home=/web/logs
>> 
>> and what is the permissions on the actual directory:
>> 
>>   ls -las d /web/logs
>> 
>> Graham
>> 
>> On 04/09/2014, at 2:58 PM, Graham Dumpleton <[email protected]> wrote:
>> 
>>> 
>>> On 04/09/2014, at 2:38 PM, Tim Arnold <[email protected]> wrote:
>>> 
>>>> Hi, I'm running a bottle application under Apache2.2.17 with mod_wsgi. My 
>>>> error log is filling up with messages like this, about 4 times per second:
>>>> [Thu Sep 04 17:20:57 2014] [crit] (45)Operation not supported: mod_wsgi 
>>>> (pid=28373): Couldn't acquire accept mutex 
>>>> '/web/logs/wsgi.17210.0.1.sock'. Shutting down daemon process.
>>>> 
>>>> This happened today after I introduced a bug into my bottle application 
>>>> and restarted Apache. I backed out the error and that's when the error 
>>>> messages started filling up. I know I must have changed something else but 
>>>> I can't see it.
>>>> 
>>>> Right now I have a cron job to empty the error log every few minutes 
>>>> because the web pages are actually delivered, although the pages are slow.
>>>> 
>>>> Here's what I've done so far:
>>>> I can run the bottle development server with no errors. 
>>>> I have rebooted the machine, but no change in behavior.
>>>> 
>>>> This is set up for Apache in httpd.conf:
>>>> LogLevel warn
>>>> WSGIPassAuthorization On
>>>> WSGIScriptAlias /service /web/home/bottle/bottle_adapter.wsgi
>>>> WSGIDaemonProcess example01 user=www group=wheel processes=5 threads=25 
>>>> home=/web/logs
>>> 
>>> Do not user user/group unless you have a need to run as a user other than 
>>> the default Apache user. Not having options will use default Apache user.
>>> 
>>> This in itself shouldn't cause the problem I don't think.
>>> 
>>>> This is the log directory (/web/logs)
>>>> -rw-------  1 root  wheel       0 Sep  4 16:53 accept.lock.17210
>>>> -rw-r--r--  1 root  wheel    4167 Sep  4 17:08 access_log
>>>> -rw-r--r--  1 root  wheel  188790 Sep  4 17:27 error_log
>>>> -rw-r--r--  1 root  wheel       6 Sep  4 16:53 httpd.pid
>>>> srwx------  1 www   wheel       0 Sep  4 16:53 wsgi.17210.0.1.sock=
>>>> 
>>>> This is from ldd:
>>>> ./mod_wsgi.so:
>>>>         libutil.so.8 => /lib/libutil.so.8 (0x800889000)
>>>>         libm.so.5 => /lib/libm.so.5 (0x800ee1000)
>>>>         libc.so.7 => /lib/libc.so.7 (0x800647000)
>>> 
>>> This indicates a separate problem. Read:
>>> 
>>> http://code.google.com/p/modwsgi/wiki/CheckingYourInstallation#Python_Shared_Library
>>> 
>>> Specifically, your Python installation wasn't installed with a shared 
>>> library.
>>> 
>>> I can't see how that would cause issue though.
>>> 
>>>> The httpd.conf syntax is okay:
>>>> sudo apachectl -t
>>>> Syntax OK
>>> 
>>> What Apache MPM are you using. You can find out using 'httpd -V' if you 
>>> don't know.
>>> 
>>>> The access log looks normal. The only thing in the error log are those 
>>>> [crit] errors.
>>>> I can't think of anything else to check. Any pointers on where to start 
>>>> looking? 
>>> 
>>> Have you enabled and or changed configuration for SELinux on your system?
>>> 
>>> What Linux distribution are you using?
>>> 
>>> Graham
>>> 
>> 
>> 
>> -- 
>> You received this message because you are subscribed to the Google Groups 
>> "modwsgi" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to [email protected].
>> To post to this group, send email to [email protected].
>> Visit this group at http://groups.google.com/group/modwsgi.
>> For more options, visit https://groups.google.com/d/optout.
> 

-- 
You received this message because you are subscribed to the Google Groups 
"modwsgi" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/modwsgi.
For more options, visit https://groups.google.com/d/optout.

Reply via email to