Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[roseway]

If you haven't already noticed, if DSLstats was running at the time when the clocks went back, then it didn't appear to notice the change of time, and the graphs will be an hour out. Restarting it will correct this. If you use the HG612-Modem-Stats co-operation feature, then this will also have been put out of synchronisation.

I know the cause of this, and I'll fix it before the clocks go forward again next spring.

[renluop]

's funny as my graphs are showing right time (10:40) and router shows no resets.

Event Log

Code: [Select]

26 Oct 2013 22:36:35 Recording started

26 Oct 2013 22:36:44 IP address is now 
27 Oct 2013 01:36:42 Average errors saved as 'errors1.txt' in snapshot directory
27 Oct 2013 01:36:42 Auto snapshots taken
27 Oct 2013 04:37:12 Average errors saved as 'errors2.txt' in snapshot directory
27 Oct 2013 04:37:12 Auto snapshots taken
27 Oct 2013 07:37:42 Average errors saved as 'errors3.txt' in snapshot directory
27 Oct 2013 07:37:42 Auto snapshots taken
27 Oct 2013 10:38:12 Average errors saved as 'errors4.txt' in snapshot directory
27 Oct 2013 10:38:12 Auto snapshots taken


[roseway]

I see, thanks for that. Perhaps it's only the Linux version which misbehaves.

[renluop]

But then it's still strange, as the official change time is 0200 hrs BST. I'd have thought that for example one should see records in BST up to 0159 and the bext one at, say, 0100 GMT. That doesn't seem to be.

Someone's stolen my hours?!

[roseway]

The way it works is that, as each sample time arrives it calculates the time when the next sample is due. It then repeatedly reads the system time from the PC and compares this with the time when the next sample is due. So assuming that your PC time was adjusted correctly (put back one hour at 02:00), it would have missed out an hour of sampling while it waited for the PC time to catch up with the time when the next sample is due.

When the clocks next go forward it will go slightly crazy unless I modify the method a bit (which I will).

[kitz]

's funny as my graphs are showing right time (10:40) and router shows no resets.

It is currently showing correct time for me too ,  although Ive no idea how it coped with the fact there would have been a duplicate time frame of results.   


27 Oct 2013 01:33:56   Auto snapshots taken
27 Oct 2013 03:34:56   Auto snapshots taken

[roseway]

although Ive no idea how it coped with the fact there would have been a duplicate time frame of results.

It would simply have stopped sampling for an hour, until the PC time caught up with when the next sample was due.

[kitz]

Thank you for the explanation eric.   Yes that makes sense