Date: Fri, 12 Aug 2016 11:41:41 -0700 From: Alfred Perlstein <alfred@freebsd.org> To: Zaphod Beeblebrox <zbeeble@gmail.com>, FreeBSD Hackers <freebsd-hackers@freebsd.org> Subject: Re: Panic not dumping to USB. Message-ID: <483f7abd-d19f-ce22-79b8-d5d5501666d0@freebsd.org> In-Reply-To: <CACpH0MeyEG1eG=ivndFFXnXNh2NocmG1xybqcFamr8McW_2UHQ@mail.gmail.com> References: <CACpH0MeyEG1eG=ivndFFXnXNh2NocmG1xybqcFamr8McW_2UHQ@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 8/12/16 10:35 AM, Zaphod Beeblebrox wrote: > My swap spaces are too small to hold some of the panic's. Either that, or > this is also happening to them. I added a 64G USB drive to the machine for > the express purpose of crash dumps. I formatted it with GPT and a label > (so that it wouldn't get mixed up with SCSI drives). The first partition > has the type freebsd-swap and the label 'crash' ... so /dev/gpt/crash is my > configured crash dump location. > > Luckily I have a serial console, so this is what's happening at crash-time: > > panic: dva_get_dsize_sync(): bad DVA 1573890:1587590144 > cpuid = 0 > KDB: stack backtrace: > #0 0xffffffff80e55fed at kdb_backtrace+0x8d > #1 0xffffffff80df76af at vpanic+0x1df > #2 0xffffffff80df74d0 at vpanic+0 > #3 0xffffffff822b8b01 at dva_get_dsize_sync+0xb1 > #4 0xffffffff822b8d9e at bp_get_dsize+0x12e > #5 0xffffffff8225a786 at dmu_tx_count_free+0x2f6 > #6 0xffffffff8225a334 at dmu_tx_hold_free+0x384 > #7 0xffffffff8223e9aa at dmu_free_long_range_impl+0xfa > #8 0xffffffff8223e850 at dmu_free_long_range+0x60 > #9 0xffffffff82312dfe at zfs_rmnode+0xbe > #10 0xffffffff822ed36f at zfs_zinactive+0xef > #11 0xffffffff82339da7 at zfs_freebsd_reclaim+0x87 > #12 0xffffffff81619894 at VOP_RECLAIM_APV+0x174 > #13 0xffffffff80f15e39 at VOP_RECLAIM+0x39 > #14 0xffffffff80f0ff99 at vgonel+0x3e9 > #15 0xffffffff80f105e1 at vrecycle+0xb1 > #16 0xffffffff8233614e at zfs_inactive+0x1ae > #17 0xffffffff82339d14 at zfs_freebsd_inactive+0x34 > Uptime: 1d5h2m19s > Dumping 10472 out of 49114 > MB:..1%..11%SESC[0mESC[2;30;40mESC[0mESC[1m^@ESC[01;0 > > ... obviously, that last bit after 11% is the console rebooting. Note that > dumping begins after Uptime. Is that normal? In this case, the 10472 > would fit on the 16G swap, but I'm not sure what the problem is. It > definitely fits on the 64G USB stick. > Last I checked there is not enough of the system running during crash to dump to USB. You may have more luck attaching an actual disk, or you might be able to use "textdumps" to get more information if your swap is too small. -Alfred
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?483f7abd-d19f-ce22-79b8-d5d5501666d0>