openSuse v9.3 (Yes, it’s old)
linux 2.6.11.4-21.17-smp i686
At some time after July 28 the entire filesystem has become read-only. It was functional on July 28, the last time I logged into the machine and edited a file. Even as the superuser when I attempt to change, say, the permissions of a file, I get:
> # chmod 777 /opt
chmod: changing permissions of `/opt': Read-only file system
jimoe666 wrote:
> At some time after July 28 the entire filesystem has become read-only.
> It was functional on July 28, the last time I logged into the machine
> and edited a file.
what file did you edit on the last time it worked as expected? did you
edit as root? (what it a system file or a file in /home ?)
did you save a copy of that file as it existed prior to your edit?
can you fall back to that pre-edit state?
why would you want to chmod 777 /opt anyway? opt is owned by root and
should stay owned by root and there is no valid reason i can think
of to mess up its permissions to allow normal uses in to muck up the
works…
is this v9.3 machine a server?
running without X?
are you directly accessing hardware through a keyboard connected to
the server, or is there a VM or Win-machine involved…
try this:
-open a terminal
-cd to /opt and issue
ls -hal
and copy paste it back to here…
and, cd to the directory/folder where exists the file you edited on 28
july and issue the same code, and copy/paste it also…
then, finally do this
cat /etc/fstab
free advice: i’ve used Linux since 1998 and can’t remember if i ever
did a chmod 777 on any folder, but i don’t think so…
All my efforts to restore the disk to some usable state were in vain (fsck, system repair, rescue disk, …). There were a large number of error messages about being unable to write to the disk. fsck said much the same thing: non-writable. mount was steadfastly at non-mountable, not even read-only.
I ended by replacing the disk drive and re-installing the OS.
On 2010-08-02 15:03, DenverD wrote:
> jimoe666 wrote:
>> I ended by replacing the disk drive and re-installing the OS.
>
> unfortunately it seems disks just don’t last like they used to…
A machine with an /opt partition running SuSE 9.3 (no openSUSE at that time)? That’s probably a
server. I guess it has been up many thousands of hours, so the disk is not a bad one.
jimoe666:
You should run SMART tests periodically on all your disks; as the 9.3 is old, you should get a new
version as source and recompile, to get the full functionality of smart[ctl|d]
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” GM (Elessar))
Carlos E. R. wrote:
> A machine with an /opt partition running SuSE 9.3 (no openSUSE at
> that time)? That’s probably a server. I guess it has been up many
> thousands of hours, so the disk is not a bad one.
that was my guess too…
and, i guess when ‘suddenly’ (after a file edit) the disk was read
only the Admin tried a MS-Standard Fix (a reboot) and THEN is when the
used-to-be-constantly-spinning-for-thousands-of-hours drive decided to not restart…
On 2010-08-03 10:04, DenverD wrote:
> Carlos E. R. wrote:
>> A machine with an /opt partition running SuSE 9.3 (no openSUSE at
>> that time)? That’s probably a server. I guess it has been up many
>> thousands of hours, so the disk is not a bad one.
>
> that was my guess too…
>
> and, i guess when ‘suddenly’ (after a file edit) the disk was read
> only the Admin tried a MS-Standard Fix (a reboot) and THEN is when the
> used-to-be-constantly-spinning-for-thousands-of-hours drive decided to
> not restart…
>
> you have seen that too, right?
Oh, yes.
It runs an fsck, which fails, then… gosh.
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” GM (Elessar))