Hi Mike,
I wonder if the retry count could be reset after a successful connection.
73 Alan M0NNB

On Mon, Jul 12, 2021 at 5:06 PM Alan Hopper <[email protected]>
wrote:

> Hi Mike,
> this is with WSJT-X connecting to the fake rigctld in SparkSDR.  It
> connects and then if a delay happens some time later (hours in real use) it
> will reconnect again but if the delay happens again it does not retry.
> Maybe be retry was planned for the initial connect? For long running
> connections it would be nice if it kept reconnecting.
> 73 Alan M0NNB
>
> On Mon, Jul 12, 2021 at 4:45 PM Black Michael via wsjt-devel <
> [email protected]> wrote:
>
>> It should be trying to reconnect twice with 3 second timeout each time.
>>
>> What rig are you using?  You could increase the timeout value  by adding
>> a switch to rigctld
>>
>> rigctld --set-conf=timeout=10000
>>
>>
>> You could also try increasing the retry in rigctld.c from 3 to maybe 4 or
>> 5 and see if that makes it happy.
>>
>>
>>         if (retcode < 0 && !RIG_IS_SOFT_ERRCODE(-retcode))
>>
>>         {
>>
>>             int retry = 3;
>>
>>
>>
>> Mike W9MDB
>>
>>
>>
>>
>> On Monday, July 12, 2021, 10:08:36 AM CDT, Alan Hopper <
>> [email protected]> wrote:
>>
>>
>> Hi all,
>> I've been doing some testing to try and make the hamlib NET rigctl
>> connection to SparkSDR as robust as possible. Very infrequently (maybe once
>> in 24hrs) responses from spark are delayed by a few seconds, This causes a
>> time out at the WSJT-X side which then auto reconnects which is great,
>> however if it happens a second time the auto retry does not seem to happen
>> (I see no attempt to reconnect) and a dialog pops up with :-
>>
>> "Hamlib error: Communication timed out
>> read_string called, rxmax=1024
>> read_string(): Timed out 3.000 seconds after 0 chars
>> rig.c(1000):rig_open return(-5) while opening connection to rig
>>
>> Timestamp: 2021-07-11T20:55:06.057Z"
>>
>> I've added an artificial delay to test and the auto retry only ever
>> appears to happen once. A manual retry does work after the second failure.
>>
>> In the ideal world spark would never send a delayed response but it is
>> often used on heavily loaded pcs running many WSJT-X instances 24/7.  The
>> delays happen in a low level socket send so I don't think I have any
>> control over it.
>>
>> I see it in 2.4.0 and 2.5.0rc3, I've not tested with any other versions.
>>
>> Is this intended behaviour or a bug?
>>
>> 73 Alan M0NNB
>>
>>
>> _______________________________________________
>> wsjt-devel mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/wsjt-devel
>> _______________________________________________
>> wsjt-devel mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/wsjt-devel
>>
>
_______________________________________________
wsjt-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/wsjt-devel

Reply via email to