[DXLog.net - Support] Log sync on lossy network

[9A5K]

Hi Pete.

Thanks for the info.
I've noticed the same at 9M0W operation, when network is not 100% reliable.

Network synch works on send, confirm method, so if confirmation isn't
received correctly, it will stop,
and after some time it will restart again (after checksum info is received).

Let me check a bit here how to change it.


On Sun, Mar 18, 2018 at 3:53 PM, Pete Lindsay <pete.g4cla at googlemail.com>
wrote:

>  I have noticed that when using a "none perfect" network DXLog will not
> synchronise large log files.
> We used DXLog @ PZ5K on two laptops for ARRL SSB with 6000 QSOs.
> The network was not ideal and connected via a Wifi router there was some
> packet loss.
>
> The two logs would not synchronize. When I got home I looked into this a
> bit more.
> If I connect to PCs and then add some packet loss I see that with a log
> with a large number of QSO it is not possible to synchronise the logs.
> My set up was this
> One PC containing 6000 QSOs.
> The second PC then starts up with an empty log.
> After only a few minutes all QSOs appear in the empty log and all is OK.
> Now repeat this with an application running on the second PC that simulates
> packet loss. I used "clumsy" and simulated 10% packet loss.
> This time the second PC receives some QSOs but soon the copying of QSOs
> stops.
> I tried disabling and then re-enabling the network in DXLog but I can see
> using a packet sniffer that the synchronisation starts again at the
> beginning of the log, so this doesn't help.
> I tried the same experiment, with 10% packet loss using Win-test on both
> laptops and the log transfer is all OK.
> Looks like we need a UDP retry mechanism if there is a loss of packets in
> the network?
>
> 73
>
> Pete, G4CLA
> _______________________________________________
> Support mailing list
> Support at dxlog.net
> http://www.dxlog.net/mailman/listinfo/support
>