Hi Are you using Windows? And is that G drive a UNC resource? And what version of Camel are you using? There has been some Windows File IO improvements in the two latests releases.
And it could be just that the file consumer cannot send the message to the broker and it keeps waiting/blocked. A thread dump should maybe help? On Tue, May 25, 2010 at 3:49 PM, Leen Toelen <[email protected]> wrote: > Hi, > > on high load I occasionally see a block of the directory based producer. > > The setup is as follows. Some service stores xml file in a directory > g:\accepted. A camel based service copies those to a queue > file://G:/accepted/ -> activemq:queue:Accepted. (5.3.2) > > On high load the system occasionally locks up, and the only trace I > see is that the .camellock file in the directory remains there > forever. Only when I restart the seconds service the lock file is > removed and the service resumes. No threads in the second service are > blocked, but only some of them are runnable, all others are waiting. I > cannot find any cause in the log files of either the second service or > activemq. Has anyone else noticed locks in the directory based > producer, or should I instead look more closely to the queueing > system? > > Regards, > Leen > -- Claus Ibsen Apache Camel Committer Author of Camel in Action: http://www.manning.com/ibsen/ Open Source Integration: http://fusesource.com Blog: http://davsclaus.blogspot.com/ Twitter: http://twitter.com/davsclaus
