jdh@bu-it.bu.edu (Jason Heirtzler) (02/01/89)
Has anyone noticed that fsck sometimes only checks the root partition
(xy0a) when you boot multi-user ? This has gotten me into trouble, when
we starting using filesystems that weren't clean.
The queer part is if you boot single-user and fsck manually with "fsck -p"
sometimes it will check all the filesystems - and sometimes it won't ! To
wit:
> b -s
[ boot stuff .. ]
using xxx buffers..
# /etc/fsck -p
[ checks only xy0a ]
# /etc/fsck -p
[ checks all partitions - on both disks! ]
# ^D
Multiuser startup in progress...
The system in question is a Sun-3/180 SunOS 3.5 with a Xylogics 451
controller and two single eagle drives. I haven't seen this on any other
of our Suns.
Jason Heirtzler
Boston University
Software Support
jdh@bu-it.bu.edu
(617) 353-2780meier@mtap.src.honeywell.com (Christopher M. Meier) (02/15/89)
jdh@bu-it.bu.edu (Jason Heirtzler) writes: > Has anyone noticed that fsck sometimes only checks the root partition > (xy0a) when you boot multi-user?... Strangely enough, yes. I noticed it for the first (and only) time 2 days ago. This was on a Sun-3/160 SunOS 3.5 with a ciprico-3200 (4drive) controller using version 1.14 of their driver, on a newly installed 1.2Meg sabre. Any ideas? @ Christopher M. Meier MN65-2300 Honeywell Systems & Research Center @ 3660 Technology Drive Mpls, MN 55418 (612) 782-7191 @ meier@SRC.Honeywell.COM | !srcsip!meier
whm@uunet.uu.net (Bill Mitchell) (02/16/89)
jdh@bu-it.bu.edu (Jason Heirtzler) writes: > Has anyone noticed that fsck sometimes only checks the root partition > (xy0a) when you boot multi-user ? This has gotten me into trouble, when > we starting using filesystems that weren't clean. I've seen something similar: Just after we brought up 4.0 an fsck -p didn't check a couple of the filesystems it was supposed to. As I recall, we verified that the associated /etc/fstab entries indicated that the filesystems should be checked. We tried to reproduce the bug, but couldn't. We recounted this to Sun and after a day they got back to us and suggested that we check the /etc/fstab entries... We asked if they'd file it as a suspected bug, but they refused. Has anyone else ever seen this? Bill Mitchell whm@sunquest.com Sunquest Information Systems sunquest!whm@arizona.edu Tucson, AZ {arizona,uunet}!sunquest!whm 602-885-7700
sklower@okeeffe.berkeley.edu (Keith Sklower) (02/28/89)
I had this happen on a 3.4 workstation which also had remote file systems mounted. One of the nfs lines in the fstab didn't have both frequency of dumping information entries. I would check the fstab very carefully, deleting comments, blank lines, running fsck single user without -p to see how far it got.