Throughout the documentation, the term “stage environment” will be used to identify the staging or test or whatever environment is used for the tests.
Set up the stage environment as a new ULS-server, you do not need any data from the production environment.
Remove all defined limits, aggregations and value retentions. Login to the ULS-server as user “uls” and issue the command:
$ u2w_interpreter /home/uls/testsuite/del_all_limits.s2w
Check that the regular ULS jobs are correctly executed by cron:
$ cd /etc/cron.d/ $ vi ulsserver
Should look like:
# delete old values 3 0 * * * uls /usr/local/bin/jobmonitor2uls -s ULS -t Delete-Old-Loggings /srv/uls/clean/mysql_delete_old_loggings >/dev/null 2>&1 # # test limits 4-59/5 * * * * uls /usr/local/bin/jobmonitor2uls -s ULS -t Test-Limits /srv/uls/ulsmeld/do_ulsmeld test_isalive test_limits test_kombilimits >/dev/null 2>&1 # # send mail reports # 3 * * * * uls /usr/local/bin/jobmonitor2uls -s ULS -t Mailreports /srv/uls/jobs/send_mailreports >/dev/null 2>&1 # # aggregation 1 5 * * * uls /usr/local/bin/jobmonitor2uls -s ULS -t Compression /srv/uls/jobs/komprimierung >/dev/null 2>&1
The ULS-server always marks the last checked value to continue with the next execution of the test scripts. So you need to reset these marks before starting the test scripts (or wait until it is executed by cron):
$ /usr/local/uls/ulsmeld/do_ulsmeld -f test_limits $ /usr/local/uls/ulsmeld/do_ulsmeld -f test_kombilimits $ /usr/local/uls/ulsmeld/do_ulsmeld -f test_isalive
That will perhaps run for a long time depending on the number of meanwhile unprocessed values
Clear the log file for the notification messages:
$ echo > /tmp/ulsserver.meldungen.out $ chmod 664 /tmp/ulsserver.meldungen.out
We made some research on what colors to use for the output of the graphs for which gnuplot is used. Found http://phrogz.net/css/distinct-colors.html very useful to play around with colors. But it took some time. Came back to these settings:
But finally, we took these 16 colors: