Crash with bpfs

From: Andrea Venturoli (ml.ventu_at_flashnet.it)
Date: 07/29/03

  • Next message: Paolo Pisati: "Re: Differences between netgraph nodes in 4.x and 5.x"
    To: freebsd-net@freebsd.org
    Date: Tue, 29 Jul 2003 11:38:57 EST
    
    

    Hello.
    I've got a production server which keeps crashing if I use bpfs to much.
    I usually only use bpf0 for dhcp server, but if I start e.g. snort and ntop, the machine will soon reboot. The same
    happens if I run tcpdump.
      
    uname -a gives:
      
    FreeBSD xxxxx.yyyyyyyy.zz 4.7-RELEASE-p9 FreeBSD 4.7-RELEASE-p9 #2: Sat Mar 22 19:25:28 CET 2003
       root@xxxxx.yyyyyyyy.zz:/usr/obj/usr/src/sys/XXXXX i386
      
      
    The hardware is an Athlon with 128MB RAM, 4 SCSI HD building two mirrored vinum volumes and two Intel NIC (one of which
    is currently unused, but was when it all began and probably will be again soon).
      
    The problem has started to show after an upgrade to 4.6 or 4.7 I believe (but I cannot recall exactly).
    I might as well try an upgrade to 4.8, but I'd rather have more insight.
      
    Following is the output of bt from gdb. Anyone can suggest where do I look next?
      
      
    #0 dumpsys () at ../../kern/kern_shutdown.c:487
    #1 0xc015b2ef in boot (howto=260) at ../../kern/kern_shutdown.c:316
    #2 0xc015b714 in poweroff_wait (junk=0xc02594cc, howto=-1071280145)
        at ../../kern/kern_shutdown.c:595
    #3 0xc021c30a in trap_fatal (frame=0xc8344abc, eva=3230566052)
        at ../../i386/i386/trap.c:974
    #4 0xc021bfdd in trap_pfault (frame=0xc8344abc, usermode=0, eva=3230566052)
        at ../../i386/i386/trap.c:867
    #5 0xc021bbc7 in trap (frame={tf_fs = 16, tf_es = 16, tf_ds = 16,
          tf_edi = 6704128, tf_esi = 1, tf_ebp = -936097008, tf_isp = -936097048,
          tf_ebx = -1065849344, tf_edx = -1066233856, tf_ecx = -1607974912,
          tf_eax = 1832612, tf_trapno = 12, tf_err = 2, tf_eip = -1072206401,
          tf_cs = 8, tf_eflags = 66066, tf_esp = -1066083072, tf_ss = -1066180606})
        at ../../i386/i386/trap.c:466
    #6 0xc0176dbf in m_getcl (how=1, type=1, flags=2)
        at ../../kern/uipc_mbuf.c:589
    #7 0xc012f2e7 in fxp_add_rfabuf (sc=0xc0a54e00, oldm=0xc074dd00)
        at ../../dev/fxp/if_fxp.c:1867
    #8 0xc012df28 in fxp_intr_body (sc=0xc0a54e00, statack=64 '@', count=-1)
        at ../../dev/fxp/if_fxp.c:1327
    #9 0xc012de3d in fxp_intr (xsc=0xc0a54e00) at ../../dev/fxp/if_fxp.c:1228
    #10 0xc0211ec2 in vec10 ()
    #11 0xc0182eb3 in biowait (bp=0xc3394184) at ../../kern/vfs_bio.c:2638
    #12 0xc018081d in bread (vp=0xc7fc00c0, blkno=360576, size=8192, cred=0x0,
        bpp=0xc8344c6c) at ../../kern/vfs_bio.c:525
    #13 0xc01cc5c2 in ffs_update (vp=0xc8256700, waitfor=0)
        at ../../ufs/ffs/ffs_inode.c:99
    #14 0xc01d5fed in ffs_fsync (ap=0xc8344cd0) at ../../ufs/ffs/ffs_vnops.c:273
    #15 0xc01d48cb in ffs_sync (mp=0xc0b99400, waitfor=2, cred=0xc0731900,
        p=0xc02b94e0) at vnode_if.h:558
    #16 0xc018b0df in sync (p=0xc02b94e0, uap=0x0) at ../../kern/vfs_syscalls.c:576
    #17 0xc015b08a in boot (howto=256) at ../../kern/kern_shutdown.c:235
    #18 0xc015b714 in poweroff_wait (junk=0xc02594cc, howto=-1071280145)
        at ../../kern/kern_shutdown.c:595
    #19 0xc021c30a in trap_fatal (frame=0xc8344df0, eva=3230566052)
        at ../../i386/i386/trap.c:974
    #20 0xc021bfdd in trap_pfault (frame=0xc8344df0, usermode=0, eva=3230566052)
        at ../../i386/i386/trap.c:867
    #21 0xc021bbc7 in trap (frame={tf_fs = 16, tf_es = 16, tf_ds = 16, tf_edi = 0,
          tf_esi = -1065680640, tf_ebp = -936096172, tf_isp = -936096228,
          tf_ebx = -1065680640, tf_edx = -1066233856, tf_ecx = -1607974912,
          tf_eax = 1832612, tf_trapno = 12, tf_err = 2, tf_eip = -1072194409,
          tf_cs = 8, tf_eflags = 66066, tf_esp = -1058056832, tf_ss = -947913056})
        at ../../i386/i386/trap.c:466
    #22 0xc0179c97 in sosend (so=0xc7c168c0, addr=0x0, uio=0xc8344ed4, top=0x0,
        control=0x0, flags=0, p=0xc77ffea0) at ../../kern/uipc_socket.c:567
    #23 0xc016d624 in soo_write (fp=0xc0ef5580, uio=0xc8344ed4, cred=0xc0c2e800,
        flags=0, p=0xc77ffea0) at ../../kern/sys_socket.c:81
    #24 0xc016a2b5 in dofilewrite (p=0xc77ffea0, fp=0xc0ef5580, fd=3,
        buf=0x8092000, nbyte=8240, offset=-1, flags=0) at ../../sys/file.h:162
    #25 0xc016a16e in write (p=0xc77ffea0, uap=0xc8344f80)
        at ../../kern/sys_generic.c:329
    #26 0xc021c5b9 in syscall2 (frame={tf_fs = 47, tf_es = 47, tf_ds = 47,
          tf_edi = 671784156, tf_esi = 8240, tf_ebp = -1077937636,
          tf_isp = -936095788, tf_ebx = 671771728, tf_edx = 671784156, tf_ecx = 3,
          tf_eax = 4, tf_trapno = 0, tf_err = 2, tf_eip = 673124360, tf_cs = 31,
          tf_eflags = 646, tf_esp = -1077937680, tf_ss = 47})
        at ../../i386/i386/trap.c:1175
    #27 0xc02109b5 in Xint0x80_syscall ()
    #28 0x8050a5c in ?? ()
    #29 0x804e065 in ?? ()
    #30 0x804d413 in ?? ()
    #31 0x804c0bd in ?? ()
      
      
      
      
      
     bye & Thanks
            av.
      
      
      

    _______________________________________________
    freebsd-net@freebsd.org mailing list
    http://lists.freebsd.org/mailman/listinfo/freebsd-net
    To unsubscribe, send any mail to "freebsd-net-unsubscribe@freebsd.org"


  • Next message: Paolo Pisati: "Re: Differences between netgraph nodes in 4.x and 5.x"

    Relevant Pages

    • Crash with bpf
      ... I've got a production server which keeps crashing if I use bpfs to much. ... I usually only use bpf0 for dhcp server, but if I start e.g. snort and ntop, the machine will soon reboot. ...
      (freebsd-questions)
    • apt-get upgrade - excessive list?
      ... u upgrade php4", I get an humbungous list of stuff to update. ... something I can't really do on a production server on a whim. ... The following packages will be REMOVED: analog apache apache-common ...
      (Debian-User)
    • Re: Upgrading to 4.9 CVS vs. binary?
      ... > have a spare to build then move over, so it will be an in place upgrade. ... try a upgrade of that one before doing it with you production server. ... If something broke on the ISO changes are that its fixed in the CVS. ... > binary upgrade when it's running the kernel and multiuser, ...
      (freebsd-questions)
    • Re: Upgrading to 4.9 CVS vs. binary?
      ... I'm currently running 4.8-p3 on my production server, ... > have a spare to build then move over, so it will be an in place upgrade. ... spare hardware to play with then: ... system behaviour that you need or some other really good reason ...
      (freebsd-questions)
    • RH 9 to FC1/FC2 and some questions
      ... I plan to upgrade a production server from RH9 to FC1....or maybe FC2 if ... should make necessary a fresh install rather than an upgrade? ... platform rather than a testbed only for ES. ... How many of you plan to use ...
      (Fedora)