monotone + trac = load

[Nathan Walp]

Daniel Atallah wrote:
> On 4/16/07, Nathan Walp <nwalp at pidgin.im> wrote:
>> Can someone who understands all the trac/monotone interaction stuff (if
>> such a person exists) take a look at homing and decide if good things
>> are happening there?  It looks like trac has 3 separate processes
>> keeping its little monotone db mirror up to date.  I can't imagine that
>> this is a good thing.
>
>
> There shouldn't be 3 processes, what makes you think that there are?
nwalp at homing:~$ ps ax
<snip>
31255 ?        Ss     0:07 /usr/sbin/apache2 -k start
31683 ?        S      0:00  \_ /usr/sbin/apache2 -k start
12679 ?        S      1:28  |   \_ /usr/bin/python 
/usr/share/trac/cgi-bin/trac.fcgi
12691 ?        S      0:41  |   |   \_ /usr/local/bin/mtn --norc 
--automate-stdio-size=1048576 --db=/home/var/trac/pidgin/monotone_sup
12680 ?        S      1:42  |   \_ /usr/bin/python 
/usr/share/trac/cgi-bin/trac.fcgi
12692 ?        S      0:41  |   |   \_ /usr/local/bin/mtn --norc 
--automate-stdio-size=1048576 --db=/home/var/trac/pidgin/monotone_sup
12681 ?        S      1:45  |   \_ /usr/bin/python 
/usr/share/trac/cgi-bin/trac.fcgi
12693 ?        S      0:50  |       \_ /usr/local/bin/mtn --norc 
--automate-stdio-size=1048576 --db=/home/var/trac/pidgin/monotone_sup
<snip>
>
> The way it is currently set up, there is a cron job that runs at :03
> and :33 that updates the mtn db that trac uses.  This is suboptimal as
> it causes errors in trac when the updates are in process but I don't
> believe there is currently a better way to do this (I've asked the
> trac-monotone author).
>
> -D
>