Museekd crashes randomly

[dafart@…]

On my archlinux x86 server I've been experiencing trouble with museekd crashing randomly after running a few hours without trouble. I have the idea that there could be a problem with some files are not properly utf8 encoded, so I'm currently trying to edit all those files, But if anyone could shine some light on this I would be very happy. Museekd version: svn revision 1208 (but the problem has been here for several months.

gdb backtrace:

(gdb) r
Starting program: /usr/bin/museekd 
[Thread debugging using libthread_db enabled]
[newnet.net.debug] 4096 file descriptors available for museekd.
[museekd.peers.warn] No valid client bind port set (range: 0 - 0).
[museekd.config.debug] Loading configuration '/home/dafart/.museekd/config.xml'.
[museekd.peers.warn] No valid client bind port set (range: 2243 - 0).
[museekd.peers.debug] Trying to bind to port 2243...
[newnet.net.debug] Listening on socket ':2243'.
[museekd.peers.debug] Listening for peers on port 2243
[newnet.net.debug] Listening on unix socket '/tmp/museekd.dafart'.
[museekd.iface.debug] Listening on '/tmp/museekd.dafart'.
[newnet.net.debug] Listening on socket '192.168.1.2:2240'.
[museekd.iface.debug] Listening on '192.168.1.2:2240'.
[museekd.up.debug] Checking if there are some uploads to start
[museekd.config.config.debug] Saving configuration to '/home/dafart/.museekd/config.xml'.
museekd: /var/abs/local/museekd-svn/src/museek+-build/./NewNet/nnbuffer.h:88: void NewNet::Buffer::seek(size_t): Assertion `n <= m_Count' failed.

Program received signal SIGABRT, Aborted.
0xb7fe1424 in __kernel_vsyscall ()
(gdb) bt
#0  0xb7fe1424 in __kernel_vsyscall ()
#1  0xb7c3fd01 in raise () from /lib/libc.so.6
#2  0xb7c4158e in abort () from /lib/libc.so.6
#3  0xb7c38638 in __assert_fail () from /lib/libc.so.6
#4  0x082428a7 in Museek::MessageProcessor::parseMessage(NewNet::ClientSocket*) ()
#5  0x081d6c0c in Museek::MessageProcessor::onDataReceived(NewNet::ClientSocket*) ()
#6  0x0828e336 in NewNet::ClientSocket::process() ()
#7  0x08292d94 in NewNet::Reactor::eventCallback(int, short, void*) ()
#8  0x08292f64 in eventCallback(int, short, void*) ()
#9  0xb7e72194 in event_base_loop () from /usr/lib/libevent-1.4.so.2
#10 0xb7e722e9 in event_loop () from /usr/lib/libevent-1.4.so.2
#11 0xb7e7230e in event_dispatch () from /usr/lib/libevent-1.4.so.2
#12 0x08292fd9 in NewNet::Reactor::run() ()
#13 0x08204ffc in main ()

[anonymous]

I changed all files to proper utf8 but the error still happends. Today I also tried for the first time using muscand, but this definitely makes museekd crash; Backtrace:

(gdb) r
Starting program: /usr/bin/museekd 
[Thread debugging using libthread_db enabled]
[newnet.net.debug] 4096 file descriptors available for museekd.
[museekd.peers.warn] No valid client bind port set (range: 0 - 0).
[museekd.config.debug] Loading configuration '/home/dafart/.museekd/config.xml'.
[museekd.peers.warn] No valid client bind port set (range: 2243 - 0).
[museekd.peers.debug] Trying to bind to port 2243...
[newnet.net.debug] Listening on socket ':2243'.
[museekd.peers.debug] Listening for peers on port 2243
[newnet.net.debug] Listening on unix socket '/tmp/museekd.dafart'.
[museekd.iface.debug] Listening on '/tmp/museekd.dafart'.
[newnet.net.debug] Listening on socket '192.168.1.2:2240'.
[museekd.iface.debug] Listening on '192.168.1.2:2240'.
[museekd.up.debug] Checking if there are some uploads to start
[museekd.config.config.debug] Saving configuration to '/home/dafart/.museekd/config.xml'.

Program received signal SIGHUP, Hangup.
0x08172aa4 in __gnu_cxx::__normal_iterator<std::string*, std::vector<std::string, std::allocator<std::string> > > std::__find<__gnu_cxx::__normal_iterator<std::string*, std::vector<std::string, std::allocator<std::string> > >, std::string>(__gnu_cxx::__normal_iterator<std::string*, std::vector<std::string, std::allocator<std::string> > >, __gnu_cxx::__normal_iterator<std::string*, std::vector<std::string, std::allocator<std::string> > >, std::string const&, std::random_access_iterator_tag) ()
(gdb) bt
#0  0x08172aa4 in __gnu_cxx::__normal_iterator<std::string*, std::vector<std::string, std::allocator<std::string> > > std::__find<__gnu_cxx::__normal_iterator<std::string*, std::vector<std::string, std::allocator<std::string> > >, std::string>(__gnu_cxx::__normal_iterator<std::string*, std::vector<std::string, std::allocator<std::string> > >, __gnu_cxx::__normal_iterator<std::string*, std::vector<std::string, std::allocator<std::string> > >, std::string const&, std::random_access_iterator_tag) ()
#1  0x0824ec3e in Museek::Museekd::isPrivileged(std::string) ()
#2  0x08259cab in Museek::UploadManager::queueLength(std::string const&, std::string const&) ()
#3  0x08136511 in Museek::IfaceManager::onUploadUpdated(Museek::Upload*) ()
#4  0x0825adda in ?? ()
#5  0x0825bf1f in Museek::Upload::setState(_TrState) ()
#6  0x0825f2ed in Museek::UploadManager::add(std::string const&, std::string const&, unsigned int const&, bool, bool) ()
#7  0x081e8dab in Museek::PeerSocket::onQueueDownloadRequested(PQueueDownload const*) ()
#8  0x081ec0a4 in Museek::PeerSocket::onMessageReceived(Museek::MessageProcessor::MessageData const*) ()
#9  0x08242796 in Museek::MessageProcessor::parseMessage(NewNet::ClientSocket*) ()
#10 0x081d6c0c in Museek::MessageProcessor::onDataReceived(NewNet::ClientSocket*) ()
#11 0x0828e336 in NewNet::ClientSocket::process() ()
#12 0x08292d94 in NewNet::Reactor::eventCallback(int, short, void*) ()
#13 0x08292f64 in eventCallback(int, short, void*) ()
#14 0xb7e72194 in event_base_loop () from /usr/lib/libevent-1.4.so.2
#15 0xb7e722e9 in event_loop () from /usr/lib/libevent-1.4.so.2
#16 0xb7e7230e in event_dispatch () from /usr/lib/libevent-1.4.so.2
#17 0x08292fd9 in NewNet::Reactor::run() ()
#18 0x08204ffc in main ()
(gdb) 

[gandalf]

Hi,

Are you using latest svn revision ?

Regards, Adam.

[dafart@…]

Yes this is revision 1208, the latest afaik.

[dafart@…]

anyone actually reading this? how alive is this project?

[dimas000]

i am having the same trouble. after couple of days running museekd randomly crashes. i ran it in screen and got nothing interesting in traceback. saving configuration, blah-blah-blah... and today i saw segfault message in the end of log. just "segmentation fault", no description, no reasons... i'm on Debian testing and using museekd 1:0.2+svn20100315.r1208-1 (as aptitude says) from Debian repo on thia site

[dimas000]

hm, museekd from original Debian testing repos (1208 too) seems to work much more stable. currently working for 5 days with no problems

[dafart]

I can only run it for about 10-20 minuts max. How large is your shared collection?

My coll is about 800 GB. Would that be related to the segfaulting?