[rt-devel] Apache memory usage

Swan Lance Lance.Swan at coates.com
Fri Jun 15 02:04:02 EDT 2001


Jesse,

I was looking at the total memory (through Webmin), but you're right the
majority of the memory usage is shared as below:

PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME COMMAND
9956 root       0   0 24448  22M 22812 S     0.0 35.8   0:02 httpd
9965 apache     0   0 24496 9424  4940 S     0.0 14.7   0:00 httpd
9961 apache     0   0 24096 8120  4588 S     0.0 12.6   0:00 httpd
9966 apache     0   0 28464 7948  2912 S     0.0 12.4   0:00 httpd
9959 apache     0   0 27416 6996  3048 S     0.0 10.9   0:00 httpd
9962 apache     0   0 22604 1808  1656 S     0.0  2.8   0:00 httpd
9960 apache     0   0 22616 1756  1604 S     0.0  2.7   0:00 httpd
9963 apache     0   0 22604 1740  1588 S     0.0  2.7   0:00 httpd
9964 apache     0   0 22604 1740  1588 S     0.0  2.7   0:00 httpd


Unfortunately, the server kind of 'snuck' into my server farm and is in fact
a spare desktop PC. Monitoring the processes under 'top', the swap usage
never went over 60MB, so another 64MB will go a long way (another 128MB even
longer!). Now that Linux has proved its worth, thanks mostly to RT 1.07, I'm
looking at a bigger server which should cope.

I look forward to the fastcgi version, but agree that low memory servers are
not the ideal so shouldn't be a priority.

Thanks for the response.

Kind regards
Lance

-----Original Message-----
From: Jesse [mailto:jesse at fsck.com]
Sent: 14 June 2001 07:29
To: Swan Lance
Cc: rt-devel at fsck.com
Subject: Re: [rt-devel] Apache memory usage


How are you reading the memory usage? You should note that a large
chunk of RT's memory usage is shared between all processes.  But yes,
mod_perl is a hog.  Running RT + other apps on a poor little 64 meg box
without a bunch of tuning to make sure there are only 2-3 RT webservers
running at a given time strikes me as a bad call. About 3 months ago,
I took my development laptop from 64M to 128M.   I _am_ working on
a FastCGI version of the web interface, which should be more low-memory
friendly, but it's not a really high priority right now.

        -j


On Thu, Jun 14, 2001 at 05:32:04PM +0200, Swan Lance wrote:
> Hi,
> 
> Not sure if I'm just doing something dumb (no comments required :)), but
> whenever I start up Apache with RT2 enabled, each httpd process starts
with
> an initial memory usage of about 28MB. If I startup without RT2 enabled,
> each httpd process starts with a much more reasonable memory usage of
8.5MB.
> 
> Since my little server only has 64MB of RAM, it's not long before it
starts
> swapping furiously and performance goes for a loop.
> 
> I have three other websites running on this server: RT 1.07, ezPublish
> (which uses PHP4) and the default static web site.
> 
> Has anyone else noticed this?
> 
> BTW, the version of RT2 that I'm using is 1.3.101.
> 
> Kind regards
> Lance
> 
> 
> 
> -----Original Message-----
> From: Jesse [mailto:jesse at fsck.com]
> Sent: 13 June 2001 10:00
> To: rt-devel at fsck.com
> Subject: [rt-devel] RT 1.3.100
> 
> 
> Minor cleanups and fixes.  At the rate we're going, I expect to announce 
> 2.0 Release Candidate 1 in about a week.
> 
> The Not-yet-supported mason_handler.fcgi (fastcgi) is fairly close
> to being a viable alternative for those of you who are getting burned
> by mod_perl.  I just need to make attachment display work.  You should
note
> that this isn't a release criterion for 2.0.0, just something to do in the

> evening when I'm bored ;)
> 
> When looking at tickets from a search, you should now see next / prev
links
> to navigate the search.
> 
>         -j
> 
> 2001-06-13 03:53  jesse
> 
> 	* Makefile:
> 
> 	Bumped version to 1.3.100
> 	
> 2001-06-13 03:52  jesse
> 
> 	* bin/mason_handler.fcgi:
> 
> 	The fastcgi handler works for everything except attachment viewing.
> I'm going to have to think a bit about how to set
> 	content type properly in a nice handler-agnostic manner.
> 	
> 2001-06-13 03:51  jesse
> 
> 	* bin/rt-mailgate:
> 
> 	Fixed a warn -> warning in mailgate's logging
> 	
> 2001-06-13 03:50  jesse
> 
> 	* lib/RT/Interface/Web.pm:
> 
> 	Adjusting spacing
> 	
> 2001-06-13 03:50  jesse
> 
> 	* webrt/Ticket/Elements/Tabs:
> 
> 	
> 	Added an error check to the new navigation bars to deal with a
> possible lack of active search
> 	
> 2001-06-12 21:28  jesse
> 
> 	* etc/config.pm:
> 
> 	Clarified required permissions for the RT Logdir
> 	
> 2001-06-12 19:12  jesse
> 
> 	* etc/config.pm, lib/RT/Queue.pm, lib/RT/Transaction.pm:
> 
> 	Cleaned up some docs in config.pm and Queue.pm
> 	Updated the Transaction Descriptions in transaction.pm to make sure
> 	that the Actor is listed.
> 	
> 2001-06-12 18:08  jesse
> 
> 	* webrt/Ticket/Elements/Tabs:
> 
> 	Added links to navigate within an existing search.
> 	
> 2001-06-08 23:57  jesse
> 
> 	* webrt/Ticket/Elements/EditKeywordSelects:
> 
> 	Made "(empty)" the same between "EditKeywordSelects" and
> "SearchByKeywordSelect"
> 	
> 2001-06-08 23:51  jesse
> 
> 	* lib/RT/Ticket.pm:
> 
> 	Ability to set owner by name when creating tickets.
> 	Ability to set timeleft when creating tickets
> 	
> -- 
> jesse reed vincent -- root at eruditorum.org -- jesse at fsck.com 
> 70EBAC90: 2A07 FC22 7DB4 42C1 9D71 0108 41A3 3FB3 70EB AC90
> 
> Any e-mail sent to the SLA will immediately become the intellectual
property
> of the SLA and the author of said message will enter into a period of
> indentured servitude which will last for a period of time no less than
seven
> years. 
> 
> _______________________________________________
> Rt-devel mailing list
> Rt-devel at lists.fsck.com
> http://lists.fsck.com/mailman/listinfo/rt-devel
> 
> _______________________________________________
> Rt-devel mailing list
> Rt-devel at lists.fsck.com
> http://lists.fsck.com/mailman/listinfo/rt-devel
> 

-- 
jesse reed vincent -- root at eruditorum.org -- jesse at fsck.com 
70EBAC90: 2A07 FC22 7DB4 42C1 9D71 0108 41A3 3FB3 70EB AC90

T\'waS br|ll1G 4|\||> 7#e sl1T#Y T0v3s D1|> gYR3 4nd Gimb at 1 1|\| 7#E \/\/A83
all |\/|1|\/|53Y W3R3 d4 60r0GR0V3s @|\||> |>4 M0MES wr47H oUTGR4b3.




More information about the Rt-devel mailing list