Please find the debug log file here. I hope this helps. https://dl.dropboxusercontent.com/u/11804892/PharoDebug.log
On my side, I'll switch to Transcript logging in the meantime to see if that is not coming from any errors in my implementation of the server. tx usman On Wed, May 21, 2014 at 5:15 PM, Sven Van Caekenberghe <[email protected]> wrote: > > On 21 May 2014, at 16:59, Usman Bhatti <[email protected]> wrote: > > > > > > > > > On Wed, May 21, 2014 at 1:37 PM, Sven Van Caekenberghe <[email protected]> > wrote: > > > > On 21 May 2014, at 13:30, Usman Bhatti <[email protected]> wrote: > > > > > Hi, > > > > > > When turning on ZincFileLogger on a Zinc server, after some time (1h > or so) I get the FileWrite error (see screenshot). I browsed through the > bugs related to Zinc in the Fogbugz but couldn't find any bug related to my > problem. I get the problem both on Mac (local machine) and Ubuntu > (DigitalOcean instance) on Pharo 3.0 (release). > > > > > > This is my code to register logging on the server events: > > > > > > server := ZnMultiThreadedServer startOn: 4001. > > > logger := ZnFileLogger onFileNamed: 'server-log.txt'. > > > server log addListener: logger. > > > > > > Any pointers? > > > > That is really weird: in the code above the file is named > 'server-log.txt' while in the screenshot the file is named > 'License-log.txt', how is that possible ? > > > > The error is that the file is closed, probably 'underneath' the logging > code, so the question is, how is that possible, who did that ? > > > > Oops. I renamed the log file name in my mail. I should have had a closer > look at the screenshot before sending it :((. So, my code should have > looked like it to match the error in the screenshot: > > > > server := ZnMultiThreadedServer startOn: 4001. > > logger := ZnFileLogger onFileNamed: 'License-log.txt'. > > server log addListener: logger. > > OK. > > > Can it be that two file (descriptors) got swapped ? > > > > Did you save the image, move it around, across machines ? > > > > I had this doubt of moving my development image to the cloud instance so > I created a script on the cloud instance to prepare my server from a > freshly downloaded pharo image. But I had the same problem. Contents from > the tail of the log file, if that can be of any help: > > > > 2014-05-20 21:01:03 308891 T GET /validate:123 200 18B 4ms > > 2014-05-20 21:01:03 308891 I Wrote a ZnResponse(200 OK > text/plain;charset=utf-8 18B) > > 2014-05-20 21:01:03 308891 D ZnUnknownHttpMethod bad request while > parsing > > 2014-05-20 21:01:03 308891 D ConnectionClosed: Cannot write data while > writing response > > 2014-05-20 21:01:03 308891 D Closing stream > > 2014-05-20 21:06:03 817717 D Wait for accept timed out > > 2014-05-20 21:11:03 817717 D Wait for accept timed out > > 2014-05-20 21:16:03 817717 D Wait for accept timed out > > I have to little information to say anything. I doubt that it is related > to Zn logging on itself, but who knows, anything is possible. It seems you > can reproduce it quite easily. Is there more in PharoDebug.log ? > > In most cases I use #logToTranscript, but with something like > > NonInteractiveTranscript file install. > > on top of my start.st script. That way I can easily write other things to > the same log file. > > Sven > > >
