Date: Fri, 02 May 2003 16:39:24 -0700 From: Kirk McKusick <mckusick@beastie.mckusick.com> To: Robert Watson <rwatson@FreeBSD.org> Cc: Scott Long <scott_long@btc.adaptec.com> Subject: Re: ffs_blkfree: freeing free block -- ps, traces, fsck log Message-ID: <200305022339.h42NdOTh011519@beastie.mckusick.com> In-Reply-To: Your message of "Fri, 02 May 2003 18:16:16 EDT." <Pine.NEB.3.96L.1030502175131.4408g-200000@fledge.watson.org>
next in thread | previous in thread | raw e-mail | index | archive | help
My last checkin was in mid-March, so I am highly suspicious that someone else has done something to create this problem. Unfortunately, this panic comes about because of something that caused the bitmaps to get trashed sometime (possibly much) earlier. The usual suspects are updates to the bitmap that are failing (e.g., the I/O is not being done), or in-memory trashing (e.g., wild pointers). Given the nature of the fsck output, it looks like bitmap updates that are not being properly done. A good quick test is to mount a filesystem, add some files to it, unmount it and run fsck to see if any discrepencies show up. As the unmount sometimes forces writes that would otherwise fail, running a background fsck on it (fsck_ffs -B /foo) will sometimes show the problem a bit better. Kirk McKusick
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200305022339.h42NdOTh011519>