From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chris Faylor To: Hans-Peter Nilsson Cc: overseers@sources.redhat.com Subject: Re: Imminent disk failure on /dev/sdd3? Date: Sat, 30 Dec 2000 06:08:00 -0000 Message-id: <20000727135302.D16416@cygnus.com> References: X-SW-Source: 2000/msg00864.html On Thu, Jul 27, 2000 at 01:31:37PM -0400, Hans-Peter Nilsson wrote: >If you look at the indexing log at >~htdigid/gcc/update-log.2000-07-27-fail, you see this line: > >cp: db.wordlist: Input/output error > >and the update is aborted (gracefully). > >I don't know about you, but to me that's the sign of a disk that needs to >be replaced. > >I think I've ruled out my previous changes from failure causes, and it's >not a disk full error (BTW, there was "lots" of disk left). I manually >re-started indexing, which succeeded, despite running in parallell with >the sourceware update (which was also successful). > >It's logical to see a beginning disk error with that copy; it's a 400M >file; the ht://Dig partition gets plenty of exersise. (And it's not >really needed; I *think* it could be a "mv" just as well). This could also be a disk in need of an fsck. I've seen this with linux occasionally. cgf From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chris Faylor To: Hans-Peter Nilsson Cc: overseers@sources.redhat.com Subject: Re: Imminent disk failure on /dev/sdd3? Date: Thu, 27 Jul 2000 10:53:00 -0000 Message-ID: <20000727135302.D16416@cygnus.com> References: X-SW-Source: 2000-q3/msg00155.html Message-ID: <20000727105300.sI93ppzsf0FDqOddOw9RbDuRSLHoKA2YZmzsJsuY0m4@z> On Thu, Jul 27, 2000 at 01:31:37PM -0400, Hans-Peter Nilsson wrote: >If you look at the indexing log at >~htdigid/gcc/update-log.2000-07-27-fail, you see this line: > >cp: db.wordlist: Input/output error > >and the update is aborted (gracefully). > >I don't know about you, but to me that's the sign of a disk that needs to >be replaced. > >I think I've ruled out my previous changes from failure causes, and it's >not a disk full error (BTW, there was "lots" of disk left). I manually >re-started indexing, which succeeded, despite running in parallell with >the sourceware update (which was also successful). > >It's logical to see a beginning disk error with that copy; it's a 400M >file; the ht://Dig partition gets plenty of exersise. (And it's not >really needed; I *think* it could be a "mv" just as well). This could also be a disk in need of an fsck. I've seen this with linux occasionally. cgf