Hello everybody,
I am testing the upgrade from RT 2.0.14 to 3.0.5 and nearly everything
is working fine. Only the database setup with rt2-to-rt3-v1.20 makes
some trouble.
In my RT2 database I have about 22000 tickets and reading them out to
dumpfiles finished with no errors. Writing these tickets back to the
new RT3 database structure stops after about 13000 tickets and Linux
hangs (only way is rebooting with a reset!).
It seems that the two active processes “dumpfile-to-rt-3.0” and "mysqld"
are eating the whole memory of my system. I can see that memory is taken
away for each ticket. Here is the last output of the top command I let
run by the way:
17:33:54 up 2:31, 2 users, load average: 2.63, 1.75, 1.52
48 processes: 45 sleeping, 3 running, 0 zombie, 0 stopped
CPU states: 40.7% user 16.6% system 0.0% nice 0.0% iowait 42.6%
idle
Mem: 643080k av, 633896k used, 9184k free, 0k shrd, 3576k
buff
601768k actv, 9984k in_d, 10212k in_c
Swap: 506036k av, 506036k used, 0k free 13792k
cached
PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME CPU
COMMAND
837 root 16 0 874M 465M 848 D 44.0 74.1 49:03 0
dumpfile-to-r
It also seems that “mysqld” crashed just a few seconds before the server
hang.
Some more information to my test system:
The OS is Redhat Linux 9.0 with Apache 2.0.40-21 and MySQL-3.23.56.
Perl is version 5.8.0-88 (all needed modules are installed).
I also tested with MySQL version 4.015 (I believe) and 4.1 - same
effect.
Is there probably a memory leak in dumpfile-to-rt3 or in any of the
Perl modules?
Thanks for any help.
Regards,
Heiko Burghardt