It would be very nice if it was possible to Meteohub to include some thing in the log to explain why a restart was done (I realise that in some cases that might have to be "Logging stopped: reason unknown!" ;) :laugh: ).logger (09.10.2009 22:45:07): station 0 (RFXCOM), wrong checksum (2a vs computed 35) for sensor model 1a89 (WGR800) in byte sequence: 44 1a 89 04 83 44 c0 00 00 00
logger (10.10.2009 04:09:06): data logger started.
logger (10.10.2009 04:09:07): connect station 0 (RFXCOM via TCP/IP).
logger (10.10.2009 04:09:07): connect station 1 (System Data via Plug-in).
logger (10.10.2009 04:09:07): connect station 2 (WMR-200 via USB HID).
logger (10.10.2009 04:09:07): connect station 3 (Plug-in via Plug-in).
logger (10.10.2009 04:09:07): connect station 4 (Plug-in via Plug-in).
logger (10.10.2009 04:09:07): connect station 5 (Vantage via TCP/IP).
logger (10.10.2009 04:09:07): station 5 (Vantage), syncing.
logger (10.10.2009 04:09:07): station 5 (Vantage), rain collector size 0.2 mm per tick.
Meteohub log: Extra message with reason for automated restar
Moderator: Mattk
-
- Platinum Boarder
- Posts: 873
- Joined: Fri Jan 25, 2008 6:27 pm
- Location: Isle of Skye, Scotland
Meteohub log: Extra message with reason for automated restar
Since adding a Davis VP2 station (with sensor data collected via TCP/IP from VirtualVP running on a PC, with VirtualVP getting data from a VP2 with serial datalogger connected to that PC) I've noticed occasional logging restarts in meteohub log, e.g.
Re:Meteohub log: Extra message with reason for automated res
I will make it a bit more self explanatory in the next update.
Core reason for aborts without notification in the log is catching an unexpected interrupt like Floating point error (division by zero), segmentation fault, etc.
I will change the code to catch that signals and to put them into the log. This might also help us to get such abortion situations solved. As data logging is handled by a child process that actually logs he data and a parent process that keeps an eye on the child process (incl. restarting this when necessary) hard stops of the client process are covered by the parent, but it would be good anyway to minimize events like these.
Core reason for aborts without notification in the log is catching an unexpected interrupt like Floating point error (division by zero), segmentation fault, etc.
I will change the code to catch that signals and to put them into the log. This might also help us to get such abortion situations solved. As data logging is handled by a child process that actually logs he data and a parent process that keeps an eye on the child process (incl. restarting this when necessary) hard stops of the client process are covered by the parent, but it would be good anyway to minimize events like these.
-
- Platinum Boarder
- Posts: 873
- Joined: Fri Jan 25, 2008 6:27 pm
- Location: Isle of Skye, Scotland
Re:Meteohub log: Extra message with reason for automated res
Thanks in anticipation for that change, and for your explanation of how the processes work together.admin wrote:I will make it a bit more self explanatory in the next update.