Date: Wed, 15 May 2002 16:51:37 -0500 (CDT) From: Gustaaf Brooijmans To: d0daqshifters@fnal.gov Cc: d0_daq_devel@fnal.gov, d0captain@fnal.gov Subject: detection of corrupted data Hi, I just put p11.06.01 on the farm, which will detect data corruption in the TFW crate. If data corruption occurs, error messages will get sent to the L3 SES, and those events will fail to pass L3. So, if data corruption occurs in every single event, everything will look normal on first sight, but the rate to tape will drop to 0 (or nearly 0 if a small fraction of the events are ok). However, in the L3 ses message board (which should be running on the top right screen of the daq console, or can be started with start_daq), you will see messages like this: v3 f() 1021405067 l3 L3_Filters 0 d0lx022 0 none none bad minor comment TFW, CAL and L2 bunch crossings do not match!!! CAL: 116 , 13058 TFW: 66 , 14138 L2: 116 , 13058 (note: there are currently also quite a few messages that can be ignored and look like this: v3 f() 1021400110 l3 L3_Filters 0 d0lx017 0 none none bad minor comment Invalid versionL2GBLHeaderData : Cannot set format version of class L2GBLHeaderData to 0: No code exists for this version ) The l3 ses logs can also be consulted in their archived form at /online/log/ses/l3 (the filename is quite obvious). During global running daq shifters should look at the l3 ses message board regularly to check for these errors, and may want to do a >grep TFW every now and then (if the number of errors is small you may miss them when they scroll by). During the daytime I will do this too. Cheers, Gustaaf