Cumulus Support

Just wait until it should have done a couple of Twitter updates, say an hour or so, then attach the cumulus.009 file.

Looks like it's done a successful update: https://twitter.com/Oswestryweather so just run it until it stops updating. It's quite likely that restarting it got it going, as you hadn't actually restarted it since the 9th August previously, which presumably did get it updating again, despite what you said earlier. If/when you hit the same problem that you hit previously, then hopefully the changes I put into 1.9.3 will stop it getting stuck this time.

This literally explains nothing ha.

I will keep an eye on this over the next few days and when it stops i will re-post to this thread, what did you change in 1.9.3 out of interest for the way Twitter works?

Rich

richardmhowell wrote:This literally explains nothing ha.

It demonstrates that restarting Cumulus was what it needed to get it updating again, so it was quite likely that it had some unexpected problem doing an update, and the update thread got stuck

what did you change in 1.9.3 out of interest for the way Twitter works?

I changed it to check each time whether the update thread was still running from a previous update, and attempt to abandon the previous update so that a new one could be done.

Ah ha i see, so now if Cumulus and Twitter have a lovers tiff it will randomly check to see if Twitter is still not talking to it and if it is kick it out of the bed so that it can carry on with its business Correct?

Yes, except that the check isn't random; it checks every time it needs to do a new update that the previous update didn't hang, and tries to kill off the previous update if it did hang, so it can carry on updating. It's rare that an update does hang; it copes with problems if everything behaves properly (e.g. it gets an error back from Twitter, or there's a timeout), but there seem to be some situations that cause the underlying communications code to hang.

Note that the third-party component I use for the debug log can cause Cumulus to crash if left running for a long time - two or three weeks - so at some point you will need to stop Cumulus, delete the debug.log file, and start Cumulus again, and eventually you should remove the 'logging=1' line from cumulus.ini while Cumulus is stopped.

Excellent, thanks for your support Steve much appreciated.

Will be sure to Donate to the ongoing Development of Cumulus.

Steve,

Just wanted to let you know, this seems to be resolved now i have been getting continious updates for the past week.

Richard