GNOME Bugzilla – Bug 165819
BeagleDaemon crashes with "... ERROR_NO_MORE_FILES"
Last modified: 2005-05-20 07:29:18 UTC
Steps to reproduce: 1. Start beagled 2. Wait a while ... 3. Voila! Stack trace: Optimizing Lucene Index at /home/willem/.beagle/FileSystemIndex/Index Optimized in 2.48s Unhandled Exception: System.IO.IOException: Win32 IO returned ERROR_NO_MORE_FILES. Path: /proc/loadavg in [0x001ea] (at /var/tmp/portage/mono-1.0.5-r3/work/mcs-1.0.5/class/corlib/System.IO/FileStream.cs:180) System.IO.FileStream:.ctor (string,System.IO.FileMode,System.IO.FileAccess,System.IO.FileShare,int,bool) in [0x0000c] (at /var/tmp/portage/mono-1.0.5-r3/work/mcs-1.0.5/class/corlib/System.IO/FileStream.cs:109) System.IO.FileStream:.ctor (string,System.IO.FileMode,System.IO.FileAccess,System.IO.FileShare) in <0x00081> (wrapper remoting-invoke-with-check) System.IO.FileStream:.ctor (string,System.IO.FileMode,System.IO.FileAccess,System.IO.FileShare) in [0x0002f] (at /home/willem/Development/beagle-0.0.5/Util/SystemInformation.cs:53) Beagle.Util.SystemInformation:CheckLoadAverage () in [0x00000] (at /home/willem/Development/beagle-0.0.5/Util/SystemInformation.cs:70) Beagle.Util.SystemInformation:get_LoadAverageOneMinute () in [0x000c4] (at /home/willem/Development/beagle-0.0.5/Util/Scheduler.cs:608) Beagle.Util.Scheduler:ComputeDelay (Beagle.Util.Scheduler/Priority,double) in [0x00172] (at /home/willem/Development/beagle-0.0.5/Util/Scheduler.cs:749) Beagle.Util.Scheduler:Worker () in <0x00053> (wrapper delegate-invoke) System.MulticastDelegate:invoke_void () Other information: Seems to be crashing in Beagle.Util.SystemInformation:CheckLoadAverage() because of an unhandled exception. It can take a while, but eventually beagled crashes on my system. I am running gentoo 2.6.9 / gnome 2.8 / mono 1.05 / dbus 0.23 / beagle 0.05
Can you attach more of a log? Also, after this error happens, can you find out the PID of the beagle process and attach the output of "ls -l /proc/<pid>/fd" ?
Created attachment 36800 [details] Log resulting in Please take a look at the end of the file. As you can see, the beagle continues running somehow, but when I search with BEST I always get zero results. I am trying to reproduce it in debug mode right now, so maybe more info later.
Created attachment 36801 [details] ls -l /proc/<pid>/fd ls -l /proc/<pid>/fd as requested
Created attachment 36802 [details] beagled --debug output (stripped somewhat) It seems to be crashing on /proc/loadavg
*** Bug 165139 has been marked as a duplicate of this bug. ***
I used to hit this every now and then, but I haven't seen it in quite a while... can anyone reproduce this bug with beagle 0.0.6.1 or better?
I am getting this exception in the latest built beagle from CVS (15th-April-2005). It starts with the following messages: Unhandled Exception: System.IO.IOException: Win32 IO returned ERROR_NO_MORE_FILES. Path: /proc/loadavg and then the messages.. Beagle WARN: System.IO.IOException: Lock obtain timed out keep coming for all the backends on the console where I have started beagled with --fg --debug options. At this stage neither best nor beagle-query works, i.e. the query comes up with no results.
Created attachment 45285 [details] beagled --fg --debug log with the Unhandled exception
Can you reproduce this crash consistently? Can we see the entire log? That fragment don't contain enough information?
Created attachment 45377 [details] Complete Log file for Beagle from .beagle/Log directory
Created attachment 45378 [details] Complete Beagle Index Helper Log file from .beagle/Log directory
Jon: I am not sure under which condition does this problem occurs, but I would try to reproduce the problem. I have attached the complete log files (Beagle and IndexHelper) before this comment.
Created attachment 45416 [details] Another example of a ERROR_NO_MORE_FILES Beagle
Created attachment 45417 [details] Another example of a ERROR_NO_MORE_FILES IndexHelper
I also get this problem on gentoo 2.6.11-r6, mono 1.06, beagle cvs from today, dbus 23.4, gnome 2.8.3 Also: it's always the same: start, wait, crash I will post the logfiles + the output of ls -l /proc/<pid>/fd Maybe this will help somehow. Note: ad-ad4 are outputs of the various pids. there are not the same (at least not according to diff)
Created attachment 45861 [details] output of the log + ls -l /proc/<pid>/fd
After I updated to mono 1.1.7 I didn't get the problem again. Beagled has been running for 2 hours straight and has now finished indexing my home directory, whereas before it stopped after 10 to 20 minutes
This is excellent news; the IO layer has been rewritten in 1.1.7 and it's possible this has been a mono bug all along.
I'm closing this as NOTGNOME. Please re-open the bug if the problem re-appears with mono 1.1.7 or better.