RT Performance Issues

Hello everybody,

I have installed RT on sun51 and am in evaluation
process. I feel RT takes a little time to load approx
6 sec for any request. Is there any way I can improve
the speed of the RT system. Why does it take some time
to load even when there is not much data. What happens
when we have lots of data in the system.
Does everyone get same behaviour.

Thanks,
Niranjan

Do you Yahoo!?
New DSL Internet Access from SBC & Yahoo!
http://sbc.yahoo.com

I had similar performance problems where RT would hang a bit between page
loads. I am running 2.0.14 on a Solaris 2.8 Netra T105.

Initial install was with 512MB of RAM, another user on this list advised that
the 512M would get chewed up fairly quickly. He was correct. With barely 50
tickets in our system things were getting a bit slow and the system was
chewing into swap space quite a bit.

I upped the RAM to 1024MB and speed improved quite nicely - except for doing
Keyword Selections. Is that database intensive work? If so, it is only db
intesive at setup time, or whenever Keywords are utilized?

I had this question on the output of ‘top’:

load averages:
CPU states: % idle, 0.0% user, % kernel, 0.0% iowait, 0.0% swap
load averages:
42 processes: 41 sleeping, 1 on cpu
CPU states: 99.8% idle, 0.0% user, 0.2% kernel, 0.0% iowait, 0.0% swap
Memory: 1024M real, 511M free, 860M swap in use, 2031M swap free

PID USERNAME THR PRI NICE SIZE RES STATE TIME CPU COMMAND
371 root 7 58 0 2392K 984K sleep 0:20 0.0% mibiisa
421 mysql 21 58 0 695M 337M sleep 0:20 0.00% mysqld
249 apache 1 20 0 22M 19M sleep 0:18 0.00% httpd
247 apache 1 58 0 19M 13M sleep 0:12 0.00% httpd
250 apache 1 52 0 19M 14M sleep 0:08 0.00% httpd
253 apache 1 58 0 18M 13M sleep 0:06 0.00% httpd
430 apache 1 58 0 19M 14M sleep 0:04 0.00% httpd
246 apache 1 58 0 19M 15M sleep 0:03 0.00% httpd
432 apache 1 58 0 19M 14M sleep 0:03 0.00% httpd
431 apache 1 51 0 18M 13M sleep 0:02 0.00% httpd
2329 dmooney1 1 59 0 2024K 1184K cpu 0:00 0.00% top
402 root 1 0 0 1136K 112K sleep 0:00 0.00% safe_mysqld
259 root 1 2 0 1752K 168K sleep 0:00 0.00% smcboot
260 root 1 12 0 1752K 80K sleep 0:00 0.00% smcboot
317 root 2 29 0 3200K 960K sleep 0:00 0.00% dmispd

What’s the deal with the ‘safe_mysqld’ - I know it’s invoked by the mysql init
script, but is the process necessary? What does it do?

regards,

Drew

I had similar performance problems where RT would hang a bit between
page loads. I am running 2.0.14 on a Solaris 2.8 Netra T105.

Initial install was with 512MB of RAM, another user on this list
advised that the 512M would get chewed up fairly quickly. He was
correct. With barely 50 tickets in our system things were getting a
bit slow and the system was chewing into swap space quite a bit.

I upped the RAM to 1024MB and speed improved quite nicely - except
for doing Keyword Selections. Is that database intensive work? If so,
it is only db intesive at setup time, or whenever Keywords are
utilized?

I had this question on the output of ‘top’:

load averages:
CPU states: % idle, 0.0% user, % kernel, 0.0% iowait, 0.0% swap
load averages:
42 processes: 41 sleeping, 1 on cpu
CPU states: 99.8% idle, 0.0% user, 0.2% kernel, 0.0% iowait, 0.0%
swap Memory: 1024M real, 511M free, 860M swap in use, 2031M swap free

PID USERNAME THR PRI NICE SIZE RES STATE TIME CPU COMMAND
371 root 7 58 0 2392K 984K sleep 0:20 0.0% mibiisa
421 mysql 21 58 0 695M 337M sleep 0:20 0.00% mysqld
249 apache 1 20 0 22M 19M sleep 0:18 0.00% httpd
247 apache 1 58 0 19M 13M sleep 0:12 0.00% httpd
250 apache 1 52 0 19M 14M sleep 0:08 0.00% httpd
253 apache 1 58 0 18M 13M sleep 0:06 0.00% httpd
430 apache 1 58 0 19M 14M sleep 0:04 0.00% httpd
246 apache 1 58 0 19M 15M sleep 0:03 0.00% httpd
432 apache 1 58 0 19M 14M sleep 0:03 0.00% httpd
431 apache 1 51 0 18M 13M sleep 0:02 0.00% httpd
2329 dmooney1 1 59 0 2024K 1184K cpu 0:00 0.00% top
402 root 1 0 0 1136K 112K sleep 0:00 0.00%
safe_mysqld 259 root 1 2 0 1752K 168K sleep 0:00
0.00% smcboot 260 root 1 12 0 1752K 80K sleep 0:00
0.00% smcboot 317 root 2 29 0 3200K 960K sleep 0:00
0.00% dmispd

What’s the deal with the ‘safe_mysqld’ - I know it’s invoked by the
mysql init script, but is the process necessary? What does it do?

safe_mysqld is a script for starting up mysqld-- also monitors and
restarts the daemon if server dies. it’s a good thing…

What’s the deal with the ‘safe_mysqld’ - I know it’s invoked by the
mysql init script, but is the process necessary? What does it do?

safe_mysqld is a script for starting up mysqld-- also monitors and
restarts the daemon if server dies. it’s a good thing…

Very definitely. :slight_smile:

I found that adding more RAM made a huge difference in RT’s performance.

from top:

5:08pm up 14 days, 9:09, 1 user, load average: 0.00, 0.00, 0.00
62 processes: 60 sleeping, 2 running, 0 zombie, 0 stopped
CPU states: 0.5% user, 0.3% system, 0.0% nice, 99.0% idle
Mem: 903184K av, 890224K used, 12960K free, 0K shrd, 52892K
buff
Swap: 530136K av, 0K used, 530136K free 278536K
cached

It was hitting swap pretty hard with 512 MB of RAM and going to 1 GB was a
big win.

Stuart Krivis Hostmaster and Purchasing Manager

APK Net, Inc. 216-241-7166 Voice
1621 Euclid Ave., Suite 1230 216-241-7522 FAX
Cleveland, OH 44115

How many tickets/users do you have… I have about 500 tickets and just
one person (me) who uses RT web interface…My box is a 1Ghz Celeron
with 192Mb of RAM… It runs fine…

I am curious because I want to know what will happen if the amount of
people/tickets gets increased.

Ticket 1 was 07-19-02 and we’re up to 493 now. About 30 users, 14 queues,
most tickets get a fair amount of comments/replies and a lot have attached
documents. I’ve also added quite a few scrips beyond the defaults. One of
the queues gets escalated every hour, so there’s activity from that too.

There’s plenty of room for hardware upgrades. New server with SMP, more
RAM, SCSI HDs, RAID… :slight_smile:

Stuart Krivis Hostmaster and Purchasing Manager

APK Net, Inc. 216-241-7166 Voice
1621 Euclid Ave., Suite 1230 216-241-7522 FAX
Cleveland, OH 44115