Summary: Problems with root filesystem writes
- From: "Aaron Taylor" <kusoneko@xxxxxxxxx>
- Date: Tue, 28 Feb 2006 12:07:16 -0800
Hi,
I received several suggestions to try FSCKing the drives. I had
already tried that to no avail. Since this is occuring on the first
boot of the system after several clean OS installations, it seems
likely that it's a larger problem.
After my initial posting, I dug out some older Solaris 10 media from
our last EduSoft update and installed with them. The exact behaviour
doesn't repeat. However, now it refuses to start X and I can't login;
literally. When I type "root" at the console login, it tells me it
can't find a bunch of PAM files and then goes back to a login prompt
before ever even reaching the password prompt. Single user mode
doesn't work and booting off a CD to a command prompt and working with
the drive shows file corruption all over the place. I also used the
media to install on an E420R and that installation worked perfectly.
At this point it's looking much more like a hardware issue so I'm
going to try some different drives and if that doesn't solve it, I'll
just send the workstation back for a replacement.
The original question:
Hi All,read/write/setuid/devices/intr/largefiles/logging/xattr/onerror=panic/dev=1d8
I've got a problem with a new Sun Blade 1000 machine I setup. It's
Solaris 10 update 1(1/06). Nobody, including root, can write to the
root filesystem. Other slices are fine. People can write to /usr,
/home, etc, etc, etc just fine.
/ is mounted read-write as shown in the "mount" command with /usr for
comparison:
/ on /pci@8,600000/SUNW,qlc@4/fp@0,0/disk@w21000004cfe6917c,0:a
read/write/setuid/devices/dev=1d80008 on Wed Dec 31 16:00:00 1969
/usr on /dev/dsk/c1t1d0s3
000b
on Sun Feb 26 14:43:17 2006installs.
However, comparing it to /usr, you can see several differences. First,
it's the actual hardware path that is mounted instead of the link in
/dev which then normally points to the hardware path in /devices.
Second, it shows the mount date as the beginning of the UNIX clock. I
would think that even if the clock was starting when the system first
got AC power, it would at least be 15-20 seconds of runtime until it
got around to mounting disks. You can also see logging is not enabled
on /. I was under that belief that UFS logging was enabled by default
on all UFS filesystems in Solaris 10.
The appropriate entries out of /etc/vfstab:
/dev/dsk/c1t1d0s0 /dev/rdsk/c1t1d0s0 / ufs 1
no -
/dev/dsk/c1t1d0s3 /dev/rdsk/c1t1d0s3 /usr ufs 1
no -
I tried reinstalling the OS and the same behaviour is occuring on both
Does anyone have any ideas what is causing this?
Thanks for the suggestions and time!
Thanks,
-Aaron Taylor
_______________________________________________
sunmanagers mailing list
sunmanagers@xxxxxxxxxxxxxxx
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
- Prev by Date: Solaris 8 - 2GB log issue
- Next by Date: Summary, is admintool gone in Solaris 10
- Previous by thread: Solaris 8 - 2GB log issue
- Next by thread: Summary, is admintool gone in Solaris 10
- Index(es):
Relevant Pages
|
|
