[solved] /sbin/rc complains about read-only file system

[yooda]

It's strange, sometimes I get no errors at all when booting, and then sometimes there are messages as:

Code: Select all

...
*Caching service dependencies...
/sbin/rc: line 492: /var/lib/init.d/softlevel: Read-only file system
mkdir: cannot create directory '/var/lib/init.d/failed': Read-only file system
/sbin/rc: line 32: /var/lib/init.d/softlevel: no such file or directory
...

Sometimes I also get messages from stuff like gawk and others, but the complaint remains the same: Read-only file system. If I reboot again, these messages sometimes disappear.

My /etc/fstab should be okay with

Code: Select all

/dev/hda3               /               reiserfs        noatime                 0 1

and I can use the system all the same, but I'm just wondering how such things sometimes happen and sometimes don't

Looks strange to me, does anybody have an idea about that?

I read stuff about corrupted filesystems possibly causing similar stuff, but I can't detect any corruptions on my reiserfs. Kernel: vanilla sources 2.6.5 and no devfs anymore, only udev.

[yooda]

I just don't get it: I shut my computer cleanly down, the next morning my box just doesn't feel like starting up normal:

Code: Select all

mkdir: cannot create directory '/var/lib/init.d/failed': Read-only file system
mkdir: cannot create directory '/var/lib/init.d/softscripts.new': Read-only file system
ln: creating symbolic link '/var/lib/init.d/softscripts.new/alsasound' to '/etc/init.d/alsasound': No such file or directory
ln: creating symbolic link '/var/lib/init.d/softscripts.new/bootmisc' to '/etc/init.d/bootmisc': No such file or directory
ln: creating symbolic link '/var/lib/init.d/softscripts.new/checkfs' to '/etc/init.d/checkfs': No such file or directory
ln: creating symbolic link '/var/lib/init.d/softscripts.new/checkroot' to '/etc/init.d/checkroot': No such file or directory
ln: creating symbolic link '/var/lib/init.d/softscripts.new/clock' to '/etc/init.d/clock': No such file or directory
ln: creating symbolic link '/var/lib/init.d/softscripts.new/consolefont' to '/etc/init.d/consolefont': No such file or directory
ln: creating symbolic link '/var/lib/init.d/softscripts.new/hostname' to '/etc/init.d/hostname': No such file or directory
ln: creating symbolic link '/var/lib/init.d/softscripts.new/hotplug' to '/etc/init.d/hotplug': No such file or directory
ln: creating symbolic link '/var/lib/init.d/softscripts.new/keymaps' to '/etc/init.d/keymaps': No such file or directory
...
...
/sbin/rc: line 32: /var/lib/init.d/softlevel: No such file or directory
* WARNING: you are stopping a boot service.
/sbin/rc: line 32: /var/lib/init.d/softlevel: No such file or directory
* WARNING: you are stopping a boot service.
/sbin/rc: line 32: /var/lib/init.d/softlevel: No such file or directory
* WARNING: you are stopping a boot service.
* Stopping input hotplugging...
* Stopping pci hotplugging...
* Stopping usb hotplugging...
/sbin/rc: line 32: /var/lib/init.d/softlevel: No such file or directory
* Bringing eth0 down...
*    Releasing DHCP lease for eth1...
*    Stopping eth1
mv: cannot stat '/var/lib/init.d/softscripts': No such file or directory
INIT: Entering runlevel: 3
mv: cannot stat '/var/lib/init.d/softscripts': No such file or directory
* Remounting root filesystem read-only (if necessary)...

* Checking root filesystem...
Reiserfs super block in block 16 on 0x303 of format 3.6 with standard journal
Blocks (total/free): 4879743/500494 by 4096 bytes
Filesystem marked as NOT cleanly umounted
Partition /dev/hda3 is mounted with write permissions, cannot check it
* Filesystem couldn't be fixed :(

Give root password for maintenance
(or type Control-D for normal startup):

# mount /dev/hda3 -o remount,ro
# reiserfsck /dev/hda3

--->
Filesystem seems mounted read-only. Skipping journal replay.
Everything successfully finished.
No corruptions found
There are on the filesystem:
        Leaves 73346
        Internal nodes 494
        Directories 29116
        Other files 308521
        Data block pointers 4552603 (255556 of them are zero)
        Safe links 0


# restart
bash: restart: command not found
# shutdown -r now

---> rebooting
---> system starts up with no errors

I've experienced this behaviour for quite a while now, sometimes everything is fine, sometimes I get some messages from my first post, and sometimes it's like described here.

So again: If somebody has an idea what might happen, please let me know!

[yooda]

hmmm, so far, things seem to work okay now with vanilla sources 2.6.7-rc1...

/edit:
definitively no problems anymore with the new kernel

[dippen]

i get this error with all 2.6.7 based ck sources.
with vanilla 2.6.7 and 2.6.7-rc1 everything works ok.
i tried to revert supermount, but it still happens.

does somebody know how to fix this WITH ck sources?

[dippen]

now i get this error with 2.6.7 too..
pls post if you know how to fix this.
thx

[noathustra]

Does anyone know what is causing this problem? Notice that when I have the problem I also have reiserfs errors of the form:

Code: Select all

There were 1 uncompleted unlinks/truncates. Completed

but the number of unlinks/truncates changes. I get this message even though I shut the system down cleanly.

Any ideas?

I am using 2.6.8-rc2-mm

[didl]

After looking through the init scripts I think I finally got rid of this
read-only issue with /var/lib/init.d during boot by
changing /etc/conf.d/rc to

Code: Select all

# Should we mount $svcdir as a tmpfs or ramfs for some speed
# increase for slower machines, or for the more extreme setups ?

svcmount="yes"

It used to say "no" before.

[bocacorazon]

The svcmount="yes" changed fixed the problem.
I don't have any idea why, though...

With my apologies for reviving this long-dormant thread, I'd appreciate your insight into this.

[cerb]

yeah, i would like to have an explanation, too. wasn't it one of the easiest thing to do for years to boot a linux system? seems not so to me now - what is all the fuzz about that ramdisks and stuff on bootup? wasn't the old process not good enough? who had to prove something here? well, you can continue to prove on my side ... all tipps followed, from yes's and no's in several configs ... vserver-sources just won't boot.

yes, emerged baselayout and/or baselayout-vsever several times ...