Date: Fri, 25 May 2007 08:01:23 -0500 From: "Roger Miranda - Digital Relay Inc." <rmiranda@digitalrelay.ca> To: freebsd-stable@freebsd.org Subject: Re: ghosthunting: machine freeze 6.2R Message-ID: <200705250801.25112.rmiranda@digitalrelay.ca> In-Reply-To: <4656A73E.9040109@vwsoft.com> References: <200705230717.l4N7HuPW010071@lurza.secnetix.de> <20070523215818.GB64723@xor.obsecurity.org> <4656A73E.9040109@vwsoft.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Friday 25 May 2007 04:07, Volker wrote: > Kris, Roger & all, > well, we replaced the broken machine (totally different hardware), > took one of the mirrored hard disks into this replacement machine > and took this replacement into production. > > Unfortunately it took less than 16 hours for this replacement > machine to also freeze. My assumption is, the freeze itself has > nothing to do with bad hardware, as it's now happening on two > different machines. This replacement doesn't have em NICs but gives > the same bad behavior (so I also think, it's not em related). As I > really do want to know what's going on, I'm now compiling a new > world + kernel with WITNESS and INVARIANTS support and see if I can > catch something. > > I'm using the following additional kernel options: > > makeoptions DEBUG=-g > options KDB > options KDB_UNATTENDED > options KDB_TRACE > options DDB > options WITNESS > options WITNESS_SKIPSPIN > options INVARIANTS > options INVARIANT_SUPPORT > options DIAGNOSTIC > options PANIC_REBOOT_WAIT_TIME=60 > Volker & Kris, Sorry for kind of hi-jacking your thread. Just hopefully we all can work together to fix this issue (if there is one). Out machine did goto in to a "hard lock". Nothing showed up on the screen or logs. I did see something from WITNESS in dmesg. See below. But this does not show up at the time of the crash. If I remember correctly this show up at boot time. We have added the following kernel options: options DDB options KDB options BREAK_TO_DEBUGGER options INVARIANTS options INVARIANTS_SUPPORT options WITNESS options DEBUG_VFS_LOCKS options DEBUG_MEMGUARD -dmesg------- softdep_waitidle: Failed to flush worklist for 0xc44ffcf8 em1: link state changed to DOWN bridge0: Ethernet address: 66:9e:c9:a6:f5:27 em0: promiscuous mode enabled em1: promiscuous mode enabled em1: link state changed to UP lock order reversal: 1st 0xc06e0820 polling (polling) @ /usr/src/sys/kern/kern_poll.c:422 2nd 0xc473070c if_bridge (if_bridge) @ /usr/src/sys/net/if_bridge.c:1976 KDB: stack backtrace: witness_checkorder(c473070c,9,c068bfbc,7b8) at witness_checkorder+0x55c _mtx_lock_flags(c473070c,0,c068bfbc,7b8,c473070c,...) at _mtx_lock_flags+0x41 bridge_input(c43d9000,c4740000,c43a828c,c43a8000,c4740000,...) at bridge_input+0x80 ether_input(c43d9000,c4740000,c43a828c,0,c0675f3c,...) at ether_input+0x122 e1000_rxeof(d43f2cbc,c0524e5d,d43f2ca8,1,5,...) at e1000_rxeof+0x19b e1000_poll(c43d9000,0,5) at e1000_poll+0x46 netisr_poll(0) at netisr_poll+0x70 swi_net(0,c4315438,c4323a80,c050a484,c4322648,...) at swi_net+0xb0 ithread_loop(c430f700,d43f2d38,c430f700,c050a484,0,...) at ithread_loop+0x1de fork_exit(c050a484,c430f700,d43f2d38) at fork_exit+0x7d fork_trampoline() at fork_trampoline+0x8 --- trap 0x1, eip = 0, esp = 0xd43f2d6c, ebp = 0 --- softdep_waitidle: Failed to flush worklist for 0xc44ffcf8 em0: link state changed to UP ----------------
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200705250801.25112.rmiranda>