[aprssig] APRS IS weirdness

Steve Dimse steve at dimse.com
Sun Feb 27 16:38:27 CST 2005

On 2/27/05 at 3:53 PM AE5PL Lists <HamLists at ametx.com> sent:

>A simple look at DG1SGE raw packets show:
>2/27/2005 21:30
>DG1SGE>APU25N,TCPIP*,qAC,DG1SGE:=4816.95N\00944.19E-Hallole ! Standby on
>DB0RZ 438,725 {UIV32N} 
>2/27/2005 21:30
>DG1SGE>APU25N,TCPIP*,qAC,DG1SGE:=4816.95N00944.19E-Hallole ! Standby on
>DB0RZ 438,725 {UIV32N} 
>Yes, this does appear to be a server problem (unless UI-View is sending
>both packets ;-).  

The smiley probably means you know this isn't the case, but here is absolute
proof, grepped from the stream:

DG1RWT>5R5S67,DB0LWL-1,DF0HHH*,qAo,DB9AZ-10:''W/l K\]Andreas QTH:Quitzoebel  Y40
DG1RWT>5R5S67,DB0LWL-1,DF0HHH*,qAo,DB9AZ-10:''W/l K]Andreas QTH:Quitzoebel  Y40

Here a backslash in the middle of a Mic-E format packet is trashed...

>Since javAPRSSrvr passes all packets as byte arrays
>(does not modify the I field in any manner), this points to other
>software which possibly considers the lines as strings.
Besides which javAPRSSrvr has been out there a long time without causing this
problem, and the error seems to be geographically restricted.

>The offending hub does not have to be inserting both packets, only the
>bad packet.  With the convoluted interconnects possible with other
>software, an unaltered packet could pass through the network just fine
>while this server with multiple upstream connects could be passing the
>mangled packet.
OK, I can see how that could happen.

>Since DG1SGE is connected directly to the APRS-IS, he could tell us
>where he is connected and that should start our search.  

Exactly the situation where I was arguing for full tracing within the q
construct, if we had that we would likely have the answer at first glance...

Enough I-told-you-so's ;-) Anyone on list from Germany that might know a
European hub where IGates in their area tend to connect?

Alternatively, we know DB9AZ-10 is connected (directly or indirectly) to the
problem server, I emailed him directly via the QRZ email address, in case he
doesn't read english anyone able to send a message in German to him
(DB9AZ at darc.de)? My experience is Bablefish et. al. is useless for technical
communication. Here is the text I sent:

It is not your fault, but packets through your IGate are being mangled upstream
of your IGate. To help us track down the problem, could you tell us what
Internet server DB9AZ-10 is connected to?

>Note that I am
>not implying that DG1SGE is causing the problem: he isn't. 

Agreed, nor is any other callsign in the postings I have made the cause of the
problem, so no one start hate-mailing innocent victims!

Steve K4HG

More information about the aprssig mailing list