Subject: Re: fsck finds DUPs
To: None <firstname.lastname@example.org>
From: Jarle Greipsland <email@example.com>
Date: 09/13/1994 15:14:44
Charles M Hannum <firstname.lastname@example.org> writes:
>> My roomates and a friend of mine have been getting a lot of DUP errors
>> when writing a large number of files and then rebooting.
> Well, you've supplied *VERY* little information.
> What kind of machine (CPU, etc.)? How much memory? How are the disks
> partitioned? Old or new format file systems? What *is* the phase of
> the moon, anyway?
I've experienced this error just recently. Approximate description
I'd just made and installed 'world'. When /etc/daily was run it complained
about a slew of 'bad file descriptors', so I decided to run fsck. I
rebooted the system and my /usr partition, an old format file system,
didn't fsck cleanly. Two 'sections' of 10-20 DUP messages each convinced
fsck to suggest that I run it manually. (As a gesture to our sensitive
readers I'll omit the passage about fsck creating a bogus lost+found
directory, and the observation that stupidity and fsck combined, i.e. 'fsck
-c2 -y /dev/rwd0f', can produce a floating point exception when the
partition has inconsistencies and that sometimes 'newfs' is a handy
System (kernel and world) built and installed just before fsck failure.
Computer: 486DX2-66, EISA motherboard, Intel brand, 8MB memory, Mach32
graphics accelerator, WD8013EBT ethernet board, aic6360 scsi controller.
Two scsi disks in the system.
Disktype: Maxtor 345MB IDE
Maxtor-7345A|NetBSD installation generated:\
/dev/wd0a / ufs rw 1 1
/dev/wd0b /tmp mfs rw,-s=32490 0 0
/dev/wd0e /var ufs rw 1 2
/dev/wd0f /usr ufs rw 1 2
Misc: There is also two scsi-disks in the system. One of them has the
'obj' tree. I haven't seen any errors on these two. They're all of the
new file system persuasion.
Phase of the moon: The Moon is Waxing Gibbous (62% of Full) (as of today :-)
"On a clear disk you can seek forever."
-- Peter O'Toole (Trinity College, Dublin)