On Sun, Jul 20, 2008 at 1:32 AM, Rob Lacey <[EMAIL PROTECTED]> wrote: > I got the same problem when I was trying to implement something in > production, having read the same Recipie. Exactly the same error in fact. I > replicated it in the rails console when trying to query the worker. Dropping > out of the console and starting again actually worked but of course doesn't > help when you want to query straight away. > > Packet::InvalidWorker > > I got the impression that it was trying to query the worker just a fraction > before the worker was ready. I couldn't gather much more information than > that. In the end I had to abandon BackgrounDRb for my project and use a > cronjob and basic queue due to time constraints, so I'd really like to know > what the problem was so that I don't give up on BDRb alltogether. I have to > say I found the rubyforge site documentation far too lacking in details so > again I look forward to see more in depth examples. > > If it is all fixed now then its a bit late for this project so maybe I'll > give it another look for the next time it BDRb seems like an appropriate > solution.
I have a fix for Linux for the above mentioned problem, looking for a fix for OS X though. Basically, it boils down to "read_nonblock()" doesn't behave in OS X, as it should. In OSX, read_nonblock , *blocks*, hence we had to switch to recv_nonblock(), which doesn't block in both Linux and OS X. But recv_nonblock is a something of a beast and doesn't work well for newly created sockets. _______________________________________________ Backgroundrb-devel mailing list [email protected] http://rubyforge.org/mailman/listinfo/backgroundrb-devel
