[aprssig] What is happening ? (long delayed packets)
mcheavens at usa.net
Thu Jul 31 11:05:15 CDT 2008
As usual you are correct Pete.
BUT, I have seen improperly formatted reverse gated packets. As I
recall I have only seen those from APRSD stations in the past. I
found the problem (years ago) by looking at the entry points for each
packets and figuring out where they came from. (I won't go into where
or why the station was doing it as it is not relevant).
As was pointed out from my post from the past the most common cause
is from I-Gates and digis that have problems.
I have seen it personally (that I was asked to diagnose and solve)
caused by four unique problems:
1. Reverse gated packets (from an APRSD station) - Station was
located on a 1000 ft tower with no RF path outside the area. When the
band would open the APRSD had a reverse gate path with 3 hops. When
the band would open, those packets that had been originally gated to
the Internet, would get reverse gated, and would end up in another
area of the state AFTER the dupe time limit as a "new" packet.
2. TNC / Radios that were not using DCD, but were relying on RF
squelch. If the squelch was loose, the TNC would que all the packets
until the buffer was full and barf all the contents in one long stream.
3. WIDE coverage TNC/Radios that hear non-stop packets. Same issue as
2 above. They can que until the buffer is full.
NOTE: I have mainly seen 2 and 3 with KPC-3's - Which (if
not using new paradigm settings) also have packet duplication and
looping due to no dupe checking if a non n-n packet was sent through it.
4. Ui-View - Memory leak which will bring the computer to a crawl and
can slow down the entry of packets to the network after the dupe
timer has expired. Solution to this is to use the scheduler to auto
restart UI-View every x number of hours or DAYS.... On some machines
I have seen this only after UI-View has run for 30-180 days. MOST
Ui-View users never see this since they run Windows and reboot more
often than this...(sad). I run two high availability server class
computers with UI-View and have seen it on both with a steady slow
down from day one until the machine in unusable after 6+ months. I
restart UI-View on those monthly now and have NO issues. (I am
migrating those locations to JavaAPRS).
I did NOT do any research on the subject packets in this discussion.
I was simply quoted from an earlier post and thought I would mention
some other issues.
At 10:05 AM 7/31/2008, you wrote:
>I may not understand what you are saying in your post. Are you
>saying that an IGate should put NOGATE in its path when gating a
>packet to RF? This is not necessary and is not supported by most
>IGates (different paths for gated packets vs. non-gated
>packets). IGates gate packets to RF as third-party packets which
>other IGates (and the javAPRSSrvr server software) will not pass in
>APRS-IS if they are properly formatted third-party packets (TCPIP in
>the third-party header).
>If this was an issue with ZL1VFO gating to APRS-IS, you would see
>different entry points for the first copy of the packets and ZL1VFO
>as the sole introducer of the dupe packets. Instead, ZL1VFO is the
>point-of-entry for all of the packets which should lead people to
>examine what is common among the dupe packets that is not present in
>the non-dupe packets. As I pointed out in my previous post, the
>single constant in the dupe packets that is not seen in the non-dupe
>packets is that the dupe packets have been digipeated by
>ZL1BQE. Therefore, any diagnostic search for the source of the
>delayed packets should start there. While ZL1BQE may not be the
>source of the problem, it is that station's digipeat that is
>consistent throughout all of the delayed packets and not seen in any
>of the "real-time" packets.
>Pete Loveall AE5PL
>pete at ae5pl dot net
> > -----Original Message-----
> > From: Mark Cheavens
> > Sent: Thursday, July 31, 2008 9:38 AM
> > To: w0jrt at yahoo.com; TAPR APRS Mailing List
> > Subject: Re: [aprssig] What is happening ? (long delayed packets)
> > Another item is if packets get reversed gated, make sure that the
> > reverse gating station uses NOGATE in any path.
> > Mark Cheavens
> > KC5EVE
> > At 09:13 AM 7/31/2008, you wrote:
> > >
> > >Some USB to serial converters can cause this. If ZL1VFO is using
> > >one of those it would be another thing to check. Some Belkin and
> > >similar cheap units have been known to be unreliable in this
> > >regard. Keyspan brand is good.
>aprssig mailing list
>aprssig at lists.tapr.org
More information about the aprssig