Welcome to the Freeway Forums

running an events based freeway

all working great except response times are slow and getting slower, eg up to 1 minute to 'confirm' an order in the admin backend, about the same to delete an order or list all orders - its starting to drive the staff potty

should it take that long for do_page_fetch('process_order'); ?

at first we thought it was the server so have moved it now to its own dedicated server, centos 5, 512 mb ram [ i know its a low power server but its got nothing to do apart from run the booking system - and the previous server was double the power and still slow..] - load averages are 0.01 (1 min) 0.00 (5 mins) 0.00 (15 mins) - thats pretty typical; orders table is 1000 rows [600kb], customers table 1500 rows [200kb], all other tables miniscule

have maxed out the standard mysql performance variables eg table_cache etc

seems wierd, any ideas?

Hi there,

We come across this type of thing from time to time. It is frustrating, I agree. Sometimes it is a matter of delays with external parties like connecting to the clickatell SMS system was slow via one IP but quicker via another fully qualified domain name (FQDN).

We have very recently optimised some of the events checkout processes that were found slow by one customer resulting in improvements from 2 mins down to 15 seconds. In Browse events we reduced the queries from 3 mins to 10 secs. What I am saying is that we are not going to avoid the issue but will help make Freeway better.

Because Freeway is large, we need give you some help to identify your slower queries and then help to optimise them. Liviu will add some info for you here about it.

Cheers,

Damo
_________________
ZacWare, Level 1 iLab, 80 Jephson Street, Toowong 4066, Queensland, Australia

There is at least one elegant solution to every real problem.

Hi there,

For slow queries , add this in my.cnf ( under [mysqld] directive ) :
log-slow-queries=/tmp/slow.log
long_query_time = 8

All queries that took more than 8 seconds to execute will be written in the log file.

Also, add a cronjob to send the slow queries log every day:
02 12 * * * /scripts/mysql/send_slow_logs.sh >/dev/null 2>&1

send_slow_logs.sh script:
#!/bin/bash
if [ -s /tmp/slow.log ] ; then
cat /tmp/slow.log|mail -s "Slow logs" info@openfreeway.org
> /tmp/slow.log
fi