httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Michael A. Pasek" <mike...@michael-pasek.com>
Subject Re: [users@httpd] Memory / Process leak
Date Wed, 25 Nov 2009 15:37:55 GMT
In users Digest 25 Nov 2009, Issue 3715, Peter N. Lewis wrote:
> Clearly, no one has any particularly helpful suggestions on what to do 
> differently.
>
> Apache, even under relatively light load, swallowed up 700+MB over two days.
>
> So the only path forward I can see is to restart apache fairly regularly 
> and delve in to switching to nginx.
>
> Thanks to those of you who tried to help.  Maybe it is specific to 
> slicehost.  Maybe its something in how I use MySQL or mod_perl.  
> Clearly this is beyond my ability to resolve, and is not some sort 
> of well known problem.  
>
> Thanks anyway,
>   Peter.

I tried to respond to Peter directly, but got:
> 451 4.4.1 reply: read error from proxy.peter.com.au.
> ... while talking to mx2.zoneedit.com.:
> >>> DATA
> <<< 554 5.7.1 <peter@stairways.com.au>: Relay access denied
> 554 5.0.0 Service unavailable
> <<< 554 5.5.1 Error: no valid recipients

Sorry to clutter up the list, but here goes....

I'm contacting you "off list", because I'm not sure if anything I have to
say would be relevant to the list.  However, when trying to figure out 
where memory leaks are, mightn't it be helpful to force a core (if
possible) of the "lost processes", then examine the core to see if 
there's any pattern in the memory which might help identify what the
lost process might have been doing ?  Even if you only do a "strings"
on the core file, you might see some data that'll give you a clue.

OK, it's a long shot, but after a few examinations, there might be 
something in common in all the "lost processes".

>Thanks to those of you who tried to help.  Maybe it is specific to slicehost.  Maybe its
something in how I use MySQL or mod_perl.  Clearly this is beyond my ability to resolve, and
is not some sort of well known problem.  

Certainly if an SQL query goes unanswered, or your perl script goes OTL
("Out To Lunch"), Apache may time out (i.e., not shown in server status),
but the child may not exit.  As you said, I don't think this is a well-
known problem, or EVERYBODY would have to be restarting their servers
periodically.  Given all the Apache servers out there, I don't think
that's happening.

>Thanks anyway,
>   Peter.

You're welcome, and -- last, but not least -- thanks for NetPresenz.   :-)

Regards,

Michael A. Pasek

---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: users-unsubscribe@httpd.apache.org
   "   from the digest: users-digest-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org


Mime
View raw message