Date: Thu, 22 Jun 1995 09:05:55 -0500 (CDT) From: Bob Willcox <bob@luke.pmr.com> To: tinguely@plains.nodak.edu (Mark Tinguely) Cc: hsu@freefall.cdrom.com, rgrimes@gndrsh.aac.dev.com, hackers@freefall.cdrom.com, phk@freefall.cdrom.com Subject: Re: dump errors Message-ID: <199506221405.JAA00498@luke.pmr.com> In-Reply-To: <199506211439.JAA17552@plains.nodak.edu> from "Mark Tinguely" at Jun 21, 95 09:39:56 am
next in thread | previous in thread | raw e-mail | index | archive | help
Mark Tinguely wrote: > > > > I'm dumping just one filesystem, an ufs filesystem. Fsck reveals no > > > errors, but the drive itself could be going bad at those sectors. > > > Fortunately, you see these types of errors too, so it may yet > > > be a software problem. > > > > In either of your cases are theses ufs filesystems larger than 2G bytes? > > I was going to ask the same question, because the lseek-s in traverse.c should > not read: > if ((int)lseek(diskfd, ((off_t)blkno << dev_bshift), 0) < 0) > > but should read: > > if (lseek(diskfd, ((off_t)blkno << dev_bshift), 0)==(off_t) -1) > > as I stated in bug report 461. I doubt this is the problem because you > would get a lot of lseek errors. > I, too, have been having these lseek errors with dump (plus some others that may be related). What I have discovered is that if I take a file system that is failing and rebuild it (cpio it off to another disk or tape, newfs, then cpio it back) dump then gets happy. I have successfully done this with two of the filesystems that I am having problems with. I still have a couple more to go :-( Note that both of these file systems were created on 1.1.5.1 and just under 2GB. -- Bob Willcox bob@luke.pmr.com (or obiwan%bob@uunet.uu.net) Austin, TX
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199506221405.JAA00498>