Hello there, Running 10.0-RELEASE, it is the second time I have a power failure and bad shutdown. It's also the second time I get a fsck failure. This time fsck has even segfault'ed. I think I will switch to ZFS. This is the log of the next boot up : http://imgur.com/rRpREKP Is it possible to automatically run fsck manually after this kind of failures? Regards, -- Demelier David
Hi, On Wed, 19 Feb 2014 11:59:49 +0100 David Demelier <demelier.david at gmail.com> wrote:> Running 10.0-RELEASE, it is the second time I have a power failure and > bad shutdown. It's also the second time I get a fsck failure. This > time fsck has even segfault'ed. > > I think I will switch to ZFS. > > This is the log of the next boot up : http://imgur.com/rRpREKP > > Is it possible to automatically run fsck manually after this kind of > failures?I thought that this got fixed over time. I have had such problems before too. After turning journaling off, I did not ran into this trouble anymore. Ok, I also invested in an UPS later. Erich
> Running 10.0-RELEASE, it is the second time I have a power failure and > bad shutdown. It's also the second time I get a fsck failure. This > time fsck has even segfault'ed.> I think I will switch to ZFS.> This is the log of the next boot up : http://imgur.com/rRpREKP> Is it possible to automatically run fsck manually after this kind of failures?> Regards,> Demelier DavidIt helps to run fsck from another disk such as a USB drive. I once was successful running fsck on a FreeBSD file system using a USB-stick installation of NetBSD. This was after NetBSD crashed with unclean shutdown. But you could do this probably at least as well using a USB installation of FreeBSD, need not have fancy stuff such as X, multimedia, etc. With OS/2, and my memory dates back to the 1990s to April 2001, it was necessary to boot from installation floppies in maintenance mode to run CHKDSK /f on drive where OS/2 was installed. But some time during the single-digit days of April 2001, CHKDSK, running automatically after a crash, not due to power outage, ran amok and trashed my installation and other hard-drive partitions too. Then I was never again able to boot OS/2 Warp in any way, always got Trap 000e or Trap 000c. I have NetBSD src and pkgsrc trees on same partition used for a FreeBSD installation, and that creates a hazard with NetBSD less stable than FreeBSD. But surely it would be good to buy a UPS, as I do, so you can shutdown gracefully instead of all-of-a-sudden. Tom
On 02/19/14 11:59, David Demelier wrote:> Hello there, > > Running 10.0-RELEASE, it is the second time I have a power failure and > bad shutdown. It's also the second time I get a fsck failure. This > time fsck has even segfault'ed. > > I think I will switch to ZFS. > > This is the log of the next boot up : http://imgur.com/rRpREKP > > Is it possible to automatically run fsck manually after this kind of failures? > > Regards, >Hi David, setting "sysctl kern.cam.ada.write_cache=0" solved this problem for me. Never happened anymore. regards, Thomas
On Wed, 19 Feb 2014, David Demelier wrote:> Running 10.0-RELEASE, it is the second time I have a power failure and > bad shutdown. It's also the second time I get a fsck failure. This > time fsck has even segfault'ed. > > I think I will switch to ZFS. > > This is the log of the next boot up : http://imgur.com/rRpREKP > > Is it possible to automatically run fsck manually after this kind of failures?Yes, but I would suggest turning off SUJ instead. Soft updates alone have been reliable for me. SUJ, not really.