tmux version? libevent version? give the full strace log for 10 seconds please
On Tue, Sep 20, 2011 at 02:05:22PM +0200, marcel partap wrote: > Hi there :) > This is a strange thing, sometimes tmux totally pwns one CPU, just > noticed it has been consuming bloody 30hours of precious 2.8GHz > processor time that should have been going to BOINC *grrr* ^^ > anyways, a crude statistical analysis after detaching all clients > > sleep 10s && killall -v strace & strace -p $(pidof tmux) 2>&1|sort|uniq > > -c|sort -n|tail > shows something like > > 256 epoll_wait(3, {{EPOLLHUP, {u32=401, u64=401}}}, 32, 11) = 1 > > 256 epoll_wait(3, {{EPOLLHUP, {u32=401, u64=401}}}, 32, 31) = 1 > > 256 epoll_wait(3, {{EPOLLHUP, {u32=401, u64=401}}}, 32, 33) = 1 > > 256 epoll_wait(3, {{EPOLLHUP, {u32=401, u64=401}}}, 32, 36) = 1 > > 259 epoll_wait(3, {{EPOLLHUP, {u32=401, u64=401}}}, 32, 10) = 1 > > 259 epoll_wait(3, {{EPOLLHUP, {u32=401, u64=401}}}, 32, 21) = 1 > > 260 epoll_wait(3, {{EPOLLHUP, {u32=401, u64=401}}}, 32, 8) = 1 > > 640 stat("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2309, ...}) = > > 0 > > 93709 ioctl(401, FIONREAD, [0]) = 0 > > 93709 readv(401, 0x7fff086aedb0, 1) = -1 EAGAIN (Resource > > temporarily unavailable) > > Backtrace of the situation (for fd=401): > > #0 0x00007fcda9125727 in ioctl () at ../sysdeps/unix/syscall-template.S:82 > > #1 0x00007fcda9614733 in get_n_bytes_readable_on_socket (fd=401) at > > buffer.c:1997 > > #2 evbuffer_read (buf=0xa4ba630, fd=401, howmuch=16384) at buffer.c:2028 > > #3 0x00007fcda9617b50 in bufferevent_readcb (fd=401, event=<optimized > > out>, arg=0xa24ee40) at bufferevent_sock.c:162 > > #4 0x00007fcda960da9c in event_process_active_single_queue > > (activeq=<optimized out>, base=<optimized out>) at event.c:1312 > > #5 event_process_active (base=<optimized out>) at event.c:1379 > > #6 event_base_loop (base=0x671e20, flags=<optimized out>) at event.c:1576 > > #7 0x0000000000422752 in server_loop () at server.c:210 > > #8 0x0000000000422bd9 in server_start () at server.c:201 > > #9 0x0000000000404951 in client_connect (path=0x66c4c0 > > "/tmp/tmux-0/default", start_server=1) at client.c:80 > > #10 0x0000000000404dc1 in client_main (argc=1, argv=0x7fff086af9e0, > > flags=1) at client.c:152 > > #11 0x0000000000429153 in main (argc=1, argv=0x7fff086af9e0) at tmux.c:395 > > > # ls /proc/$(pidof tmux)/fd/401 > > lrwx------ 1 root root 64 2011-09-20 13:38 /proc/8196/fd/401 -> > > socket:[198452054] > > What is going on? Why ch0ke on 401? Will try if previously send > tmux-fix-race-condition-check-proc-fd.patch fixes it.. > Wanted to rebo0t my box anyways. http://tinyurl.com/linux31 C: > #greetings earthlings > > ------------------------------------------------------------------------------ > All the data continuously generated in your IT infrastructure contains a > definitive record of customers, application performance, security > threats, fraudulent activity and more. Splunk takes this data and makes > sense of it. Business sense. IT sense. Common sense. > http://p.sf.net/sfu/splunk-d2dcopy1 > _______________________________________________ > tmux-users mailing list > tmux-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/tmux-users ------------------------------------------------------------------------------ All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2dcopy1 _______________________________________________ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users