[WT-support] Network problem, explosion of duplicates

[Tõnno Vähk]

Tnx Bob,

 

yes, it is a very serious issue especially as the reason is unclear and I
have never heard anything like it.

 

WT version 4.9.1 on all computers.

All wired through Router in simple Ethernet network.

No remote connections or anything else unsual. The same setup we have had
for a while and nothing was really changed apart from running CQWW SSB MS
for the first time.

 

I don’t believe in RF being an issue also as nothing was changed and it has
been RF free, can’t rule out though.

 

I think there is some conflict in logging QSOs and network synchronization.

 

Does anyone know what causes the „double QSO identifier, network will be
disabled“ error? This might help to find the issue. Can it be that two QSOs
are logged in two computers at exactly the same moment? What is the QSO
identifier? 

 

What causes the UDP packets 10035 (?) error, what does that mean?

 

tonno

 

From: support-bounces at f5mzn.org [mailto:support-bounces at f5mzn.org] On Behalf
Of Bob Wilson, N6TV
Sent: Friday, November 02, 2012 4:52 PM
To: support at win-test.com
Subject: Re: [WT-support] Network problem, explosion of duplicates

 

Sorry to hear this report.  This does sound like a serious bug caused by a
bad network connection.  But I don't recall hearing other reports of the
same thing happening anywhere else.

 

Which version of Win-Test?

All computers running exactly the same version?

All wired Ethernet or were some computers using Wireless (WiFi) or cell
phone modem cards (Huawei)?

Any remote (distant shack) connections via wtTunnel or equivalent?

 

The text commands SYNC/NOSYNC will enable/disable network synchronization on
any specific computer.


73,

Bob, N6TV

 

On Fri, Nov 2, 2012 at 7:41 AM, Tõnno Vähk <tonno.vahk at gmail.com> wrote:

Something strange and very troubling happened at ES9C MS this weekend. I was
not there so I have tried to put the picture together. Maybe someone can
figure it out...

 

5 computers/stations were in the network. At some point error messages „Lost
UDP packets“ and I guess number 10035 (or 10034?) appeared. Also error
message „double QSO identified, network will be disabled“. After one of
those messages at one computer when RELOAD was done to restore network the
last QSO by that station was multiplied by thousands of times and
synchronized to other computers!

 

So I believe one QSO appeared over 90 000 (!) times (number of dupes
exploding thus in summary window) and the whole network and all computers
got very slow till Win-Test became totally not responsive.

 

In general 100,000 QSOs could be OK in log (I saw 200,000 was ok on Ducie)
so in addition to just QSO number something else severely loaded the network
and computers, some continuous syncronization of something or I don’t know.
It took a while now for QSOs to reach other computers (before network
totally froze).

 

The only way to get back to normal was to export Cabrillo log, clean it,
delete the old log and import the cabrillo back!

 

When doing Cabrillo it did not export all the QSOs  - is there a limit on
100 000 or so?? Last hour was gone.

 

Also, a new row appeared into Summary window before Band row with big
numbers. This column has not been shown anywhere in manual.

 

We have had this „double QSO identified, network will be disabled“ often
before. I would say on average 5 times in every 48 hour multi op operation
on some computer. But always RELOAD has restored the network and all fine.

 

Can anyone have any idea how this excessive multiplication of one QSO and
total paralyzing of the network can happen? What can be the reason?? It
happened two times during this contest and ruined any chances to compete for
top in EU. The thing is a mess but I tried to write everything I heard...

 

tonno

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.f5mzn.org/pipermail/support/attachments/20121102/4b5c3a3d/attachment.html>