The log file on the TDX and the TDH800 is constructed as a "Single file Rotation Log" This means, that the log overwrites itself, when it reaches a maximum size of ~10MB. The start and end is indicated, as below with ### LOG CYCLE STARTS HERE ### and ### LOG CYCLE ENDS HERE ###:
00001: ### LOG CYCLE STARTS HERE ### 00002: [Thu Jan 01 00:00:13 1970] INFO - CLTdxUnitInitiator: SystemWatchdog callback registered for HWWatchdog 00003: [Thu Jan 01 00:00:13 1970] INFO - CLTdxUnitInitiator: System mains was removed 00004: [Thu Jan 01 00:00:13 1970] INFO - CLTriosThread: Starting thread GTSMI_async_event_deliver ptr 0x216FBE70 ID 10 ... 00612: [Fri Jan 27 07:58:58 2017] ERROR - Backend1: [Thu Jan 01 00:00:55 1970] CLQAMManager: Systime set exceeds threshold. Current Thu Jan 01 00:00:55 1970 new Fri Jan 27 07:58:58 2017 ... 00708: [Fri Jan 27 08:00:12 2017] INFO - Backend5: [Fri Jan 27 08:00:11 2017] CLWatchDogController: WatchDogController is now enabled 00709: ### LOG CYCLE ENDS HERE ###
Each line has a timestamp. The time is only correct, if the TDX/TDH800 is able to get the time from a configured frontend or NTP server and is shown in GMT +0.
The rotation is - as mentioned - not given by time, but on file size. When the file reaches the limit, the system starts all over, logging from the top of the file.
If there are ie. many errors from QAM backends. This could be Clipcounts cause by a to high payload on the MUX's. Then the log file will quickly be filled with these errors:
07428: [Wed Jan 25 20:00:24 2017] ERROR - CLTdxSwitchManager: Overloaded output 2 on backend 1. Clipcount: 53 07429: [Wed Jan 25 20:00:24 2017] ERROR - CLUnitErrorHandler: Backend 1 has errors: 0x0000000000002000 07430: [Wed Jan 25 20:00:26 2017] DBG WARN - CLSystemErrorHandler: Unit 0 has errors: 0x0000000000000400
Situations like this, can shorten the log files time period to a day or even less. With this in mind, errors should quickly be fixed.
Note: The log file is not deleted upon reboot.