Return to BSD News archive
Path: sserve!manuel!munnari.oz.au!mips!mips!sdd.hp.com!uakari.primate.wisc.edu!ames!agate!soda.berkeley.edu!wjolitz From: wjolitz@soda.berkeley.edu (William F. Jolitz) Newsgroups: comp.unix.bsd Subject: Re: Help! Bad Spot on Disc?? Message-ID: <14ju5iINNa1u@agate.berkeley.edu> Date: 22 Jul 92 15:17:06 GMT References: <NEAL.92Jul22092722@neal.ctd.comsat.com> Organization: U.C. Berkeley, CS Undergraduate Association Lines: 41 NNTP-Posting-Host: soda.berkeley.edu In article <NEAL.92Jul22092722@neal.ctd.comsat.com> neal@neal.ctd.comsat.com (Neal Becker) writes: >Try #1) After loadfd bin01, extract reports "bin01.11 corrupt". I >renamed bin01.11 to junk, and loaded another copy of bin01.11. Now >extract succeeds. > Do the two files compare (cmp) identical? >When fsck -p was run, it reported "unknown inode type...". When fsck >run I cleared several of these, but each time fsck is run again the >same errors are reported! > Please send me details (e.g. create a script file and send it). I noticed a problem on initial reboot with socket inode types, but this vanished after the first fsck and did not perpetuate (I assume that the binary dump has some unecessary socket inodes on it that cpio is not recreating properly, thus a nuisance). >Looks like a bad spot, right? But why? Doesn't the software >recognize the sectors already marked bad on the disk? Currently, the system does not do the correct thing with fresh badblocks, partially due to new work in the wd driver that alters the way that retries work. > ... >problem. This message keeps showing up (a lot): > swap_pager_finish: clean of page xxx failed > >Sounds nasty. It's not. It means that the VM system saw that the page was modified after an attempted write to disk (it was dirtied by a program accessing it). For the moment, consider it a soft error. Much more is involved here, and I especially want to here if this happens more than sporadically on systems with > 4MB of RAM, but for the moment ignore this diagnostic. Don't worry, the sky is not falling (yet) ! Bill.