| Re: [paho-dev] Paho C 1.3.0 reconnection issue |
_______________________________________________Hi Milan,
I've responded on the issue. I tried to reproduce on MacOS and Ubuntu, but haven't succeeded so far. I'm probably missing some essential aspect, so if you have a test program or fuller trace, that might help.
Ian
On 05/03/2019 15:24, Milan Tucic wrote:
Hello Ian,
I appreciate your work on the project and I am aware of load you are currently facing. That is why I am not insisting on the issue. On the other hand, a feedback/label/milestone would be a good starting point.
Regarding your question, I can't confirm for TCP. During the evaluation period, I had problem on TCP client also, but with much more complex system (few localhost clients, broker and bridges).On the isolated test env, I can't reproduce it (tried again today). There is also high number of posts on connect semaphore, but first connect (with broker up and running) succeeds.
Thanks,Milan
On Tue, Mar 5, 2019, 12:46 Ian Craggs <icraggs@xxxxxxxxxxxxxxxxxxxxxxx> wrote:
_______________________________________________Hi Milan,
I've been working on the Java client service release as well as the implementation of the MQTT V5 support in the Java client. (For the moment, I also have the _javascript_ and embedded C clients to look after, although unfortunately I haven't been able to focus on those either.) In the last two weeks I've been taking some of my unused vacation after getting the Java service release out. I was going to get back to looking at the C client as soon as I could, unfortunately there was/is more work involved in the Java client updates than I was expecting, so I felt I had to concentrate on that to at least achieve a service release, as that had been long overdue. A colleague of mine was originally working on it, but moved on before finishing, so I had to take over.
The description of the issue in 604 is 'semaphore posted too many times'. This sounds like an inconvenience, but perhaps not a high priority issue. I think it's better to start with a description of the external effects, like that below ('reconnect can take up to 25 mins') rather than the internal behaviour, as then it's easier to prioritize.
One question I have is, does this happen if the connection is not using TLS?
Thanks
Ian
On 05/03/2019 09:35, Milan Tucic wrote:
Hello community,Is there anyone who can try to reproduce issue with reconnection on Paho C 1.3.0? In case of a permanent connection, once it is lost, next successful connect will be postponed depending on number of signals sent to connect semaphore. In my case it was more then 50 connect retries. Depending on period, loss takes a lot of time (e.g. 50 x 30s = 25min).
The same can't be reproduced with Paho 1.2.0.
It is posted to project's github, but no word for more then one month. Link: https://github.com/eclipse/paho.mqtt.c/issues/604.
If there something more to clarify, just ask.
Thank you,Milan
_______________________________________________ paho-dev mailing list paho-dev@xxxxxxxxxxx To change your delivery options, retrieve your password, or unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/paho-dev-- Ian Craggs icraggs@xxxxxxxxxx IBM United Kingdom Eclipse Paho Project Lead & Mosquitto Committer
paho-dev mailing list
paho-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://www.eclipse.org/mailman/listinfo/paho-dev
_______________________________________________ paho-dev mailing list paho-dev@xxxxxxxxxxx To change your delivery options, retrieve your password, or unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/paho-dev-- Ian Craggs icraggs@xxxxxxxxxx IBM United Kingdom Eclipse Paho Project Lead & Mosquitto Committer
paho-dev mailing list
paho-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://www.eclipse.org/mailman/listinfo/paho-dev