All messages are in the process of being standardized. The date and class of the message are always indicated. There are 5 classes of message:
In case a system error occured, the message will indicate the system error at the end of the message, usually after a colon. Most of the messages will appear in following format:
Mon DD HH:MM:SS -?- blah blah blah...
May 22 12:10:29 -I- Received .opcl, .ptrs and .evts data for file 04070002. May 22 11:38:52 -W- Error sending file //oppu6/disk2/data/ptrs/0406806C.ptrsIn addition, on the ROPE message screen, the following format will be observed:
Mon DD HH:MM:SS host process -?- blah blah blah...
Jun 05 03:52:05 a-oppu3 Rope.2 -I- Closing down in order to switch_mode OFFLINE May 28 22:18:49 a-oppo RopeMum.2 -W- (17) Rope died on FZIN, run 4104, trig 8904. May 28 22:42:08 a-oppu3 FZiner.0 -W- (20) Failed to link rope_out.temp to /rope_logs/04104.0: File exists May 31 04:55:04 a-oppu17 Collect -E- Timing out /rope_outputs/16868D4114As noted previously, the log transcript files are stored in /tmp/PitLogs. As usual, to 'cd' to the /tmp directory on another node, specify the full context-dependent name:
The messaging for ROPE engines, the FilterClient and the OpcalExtract process are all done using the nMsg() function in the networking library. Each of these function groups have their own MsgLog processes to which they send their messages. All ROPE engines send their messages to the RopeMsgLog process running on //oppo2, and FilterClient (running on the Filter) and OpcalExtract (running on apoc) send their messages respectively to the FilterMsgLog and the OpcalMsgLog process running on //oppu5.