patrick.thomas@LIGO.ORG - posted 16:49, Wednesday 30 April 2014 - last comment - 17:32, Wednesday 30 April 2014(11656)
Conlog running again
Conlog is back up and running and monitoring the channel list that Dave generated yesterday. I disabled the crontask that logs the disk usage and number of database rows because I know that the latter is a long query. I also disabled the MySQL binary logging that is used to report the frequently changing channels. I'm not sure if this has helped, the CPU usage for the conlog task is still at 100%. There is also a high I/O wait. Attached is a plot of the queue size over the last two days. I stopped conlog when I found it high and plateaued yesterday afternoon. The queue size seems to take a while to come down after startup now, but it remains low afterwards.
Non-image files attached to this report
Comments related to this report
I edited the scripts used to create the conlog channel list to make them site independent and put them under SVN control in userapps/release/cds/common/scripts (conlog_create_inlcude_channel_list.py and conlog_create_pv_list.bsh).
I modified the include-file-generator to remove the monitoring of the momentary SW1 and SW2 on filtermodules as these do not give the status of what is being switched. I replaced them with the SWSTAT epics pv which unambiguously gives the switch settings.
so with about 7448 filter modules in H1, the conlog channel number decremented by this amount (take out SW1,SW2 and add SWSTAT).