[Dxspider-support] missing spots in RBN view DXspider Mojo

[Dirk Koopman]

There are some very significant differences between build 328 and 331. 
Build 328 shows more repeats than it should and this may be the reason 
why you saw *N7ET/DU7 *twice - without a + character and less than 60 
minutes apart.

Everybody on the Mojo should update to the latest build of 331 - 
especially if they are running on >= build 326 and < build 331.

It is quite difficult to see what is going on unless one is prepared 
open the logging to record all the RBN input into the debug file. It 
happens that up until about a week ago, my test node GB7TLH was running 
for about 3 weeks or so in 'unset/debug nologchan' mode - which is what 
you have to do unless you have changed the defaults in DXVars.pm. This 
results in about 1.6GB of logging. But I have been having some weird 
computer freezes and so have not been able to collect the data for the 
time period you are looking at. So I cannot say, for certain, that this 
is the reason you saw this call on one node and not the other.

It is also likely that the mechanism that was not working correctly in 
build 328 might account for the different Z: lists. But this isn't 
something I have investigated fully (yet).

73 Dirk G1TLH

On 30/08/2020 11:31, dd5xx--- via Dxspider-support wrote:
> Hello everyone,
> I am active in CW most of the time and yesterday I found out the 
> following by chance. I call CQ on different bands/frequencies and get 
> registered with at least 2 RBN skimmers when I look at 
> reversebeacon.net. At the same time I was connected via Telnet on the 
> DXspider Mojo node S50CLX and GB7DJK and used exactly the same 
> settings and filters. However, my spot was NOT registered on either or 
> sometimes both those Mojo DXspider nodes, even though at least two RBN 
> skimmers registered my call on the RBN. This happened yesterday 
> evening a lot of times, whereas today tests worked better and I missed 
> my spots only one time when watching the output of both windows. Can 
> it be possible that due to the CW contest the nodes or network were 
> overloaded somehow? But in my understanding the DXspider mojo nodes 
> with RBN funcionality should show my CQ call because it was heard by 
> minimum two RBN skimmers around the world.
> this prompted me to do more tests and compare the two Mojo nodes side 
> by side and parallel. The used versions of these two nodes are as 
> follows...
> _s50clx.infrax.si:41112_
> DX Spider Cluster version 1.57 (build*328* git: mojo/6ad88bbf[r]) on Linux
> _gb7djk.dxcluster.net:7300_
> DX Spider Cluster version 1.57 (build*331* git: mojo/d745902[r]) on Linux
> on both nodes I executed initially those commands:
> set/skimmer cw
> reject/spots all
> accept/rbn all
> I ensured, that the filters are exactly the same on both nodes...
> show/filter
> DL0ABCD : spots
>  filter1 reject all
> DL0ABCD : rbn
>  filter1 accept all
> DL0ABCD de *S50CLX* 30-Aug-2020 1018Z dxspider >
> sh/filter default
> No filters defined for DL0ABCD
> show/filter
> DL0ABCD : spots
>  filter1 reject all
> DL0ABCD : rbn
>  filter1 accept all
> DL0ABCD de *GB7DJK* 30-Aug-2020 1018Z dxspider >
> sh/filter default
> No filters defined for DL0ABCD
> I've made the following experience:
> [*1*] actually the RBN with the lowest SNR is always shown, at least I 
> took it from the Mojo/RBN description. In general this is true and 
> works fine but sometimes I realized a SNR mismatch when comparing the 
> two nodes. Attached a screenshot as example with a SNR comparison. The 
> green rectangle is showing a correct example, the yellow rectangle is 
> illustrating the different spot where SNR differs. Nothing really bad, 
> just want to mention that.
> [*2*] but what caused me more headaches was the fact that spots are 
> generally missing and don't even appear in the output. I have added 
> two example screenshots, ex1 and ex2. On ex1 screenshot all the blue 
> marked spots of S50CLX where not shown at all on GB7DJK node. And 
> vice-versa the green marked spots on GB7DJK node did not appear on 
> S50CLX node. I have triple checked the filter settings, they were 
> exactly the same. On ex2 the spot of *N7ET/DU7* appeard two times on 
> S50CLX node when I watched the window but it never appeared on GB7DJK 
> node.
> Am I doing something wrong in my comparison, did I miss any important 
> information when doing this comparison? Hope anyone can reveal some 
> information. Looking forward to your comments.
> 73
> Saki, DD5XX

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.tobit.co.uk/pipermail/dxspider-support/attachments/20200830/1ec590c0/attachment.htm>