I had a lot of updates through Yast last night. Today, when I started up the PC, I got some messages. Among them:
/dev/sda7: Unexpected Inconsistency; Run fsck Manually
(i.e. without -a or -p options)
Loading console font lat9w-16.psfu -m trivial G0:loadable
Loading Keymap i386/qwerty/us.map.gz
Start Unicode mode
fsck failed for at least one file system (not/)
Please repair manually and reboot
The root file system is already mounted read-write...
So it prompted me for the root password. Typed that in. Next prompt, I typed in fsck. Then I got a message that I will SEVERELY damage my file system if I go through with fsck, then I said “no.” It went through some other checks complaining of a “short” read in sda7 (my home dir.)
I decided to bug out with ctrl-D and went in through failsafe mode. Everything seemed to run normally, so I went back through the regular desktop. Everything ran normally, but just to be safe…
How do I check the file system and correct it if there are errors? Thanks in advance!
On Thu, 01 Mar 2012 20:56:03 +0000, HealingMindNOS wrote:
> So it prompted me for the root password. Typed that in. Next prompt, I
> typed in fsck. Then I got a message that I will SEVERELY damage my file
> system if I go through with fsck, then I said “no.” It went through some
> other checks complaining of a “short” read in sda7 (my home dir.)
>
> I decided to bug out with ctrl-D and went in through failsafe mode.
> Everything seemed to run normally, so I went back through the regular
> desktop. Everything ran normally, but just to be safe…
>
> How do I check the file system and correct it if there are errors?
> Thanks in advance!
You need to unmount the filesystem before running fsck - fsck should not
be used on a mounted filesystem (which is why you got the warning you
did).
Try (in single user mode):
umount /dev/sda7
fsck /dev/sda7
Note that if your home partition is also your root partition, this will
be more involved because fsck is on the root partition, so the partition
needs to be mounted.
Am 01.03.2012 22:19, schrieb Jim Henderson:
> Try (in single user mode):
>
> umount /dev/sda7
> fsck /dev/sda7
>
> Note that if your home partition is also your root partition, this will
> be more involved because fsck is on the root partition, so the partition
> needs to be mounted.
>
> Jim
>
It is enough if the root partition is remounted readonly in the last
case, you will not get the warning and the fsck is a safe operation if
it is readonly.
mount -r -o remount /
as root, but double check the command, I cannot test it right now.
–
PC: oS 11.4 (dual boot 12.1) 64 bit | Intel Core i7-2600@3.40GHz | KDE
4.6.0 | GeForce GT 420 | 16GB Ram
Eee PC 1201n: oS 11.4 64 bit | Intel Atom 330@1.60GHz | KDE 4.8.0 |
nVidia ION | 3GB Ram
On Thu, 01 Mar 2012 21:36:32 +0000, Martin Helm wrote:
> Am 01.03.2012 22:19, schrieb Jim Henderson:
>> Try (in single user mode):
>>
>> umount /dev/sda7 fsck /dev/sda7
>>
>> Note that if your home partition is also your root partition, this will
>> be more involved because fsck is on the root partition, so the
>> partition needs to be mounted.
>>
>> Jim
>>
> It is enough if the root partition is remounted readonly in the last
> case, you will not get the warning and the fsck is a safe operation if
> it is readonly.
>
> mount -r -o remount /
>
> as root, but double check the command, I cannot test it right now.
-o ro,remount
But I had recalled that at least once I’d done that and the filesystem
either didn’t remount as RO or I got the warning anyways. Might’ve been
some time ago, or I might be misremembering.
Am 01.03.2012 22:44, schrieb Jim Henderson:
> But I had recalled that at least once I’d done that and the filesystem
> either didn’t remount as RO or I got the warning anyways. Might’ve been
> some time ago, or I might be misremembering.
>
> Jim
>
>
You are right I tried it now, it gives the message (while I do not
understand the reason, because it makes no sense).
So the best way is to do it from a live CD.
–
PC: oS 11.4 (dual boot 12.1) 64 bit | Intel Core i7-2600@3.40GHz | KDE
4.6.0 | GeForce GT 420 | 16GB Ram
Eee PC 1201n: oS 11.4 64 bit | Intel Atom 330@1.60GHz | KDE 4.8.0 |
nVidia ION | 3GB Ram
On Thu, 01 Mar 2012 22:02:04 +0000, Martin Helm wrote:
> Am 01.03.2012 22:44, schrieb Jim Henderson:
>> But I had recalled that at least once I’d done that and the filesystem
>> either didn’t remount as RO or I got the warning anyways. Might’ve
>> been some time ago, or I might be misremembering.
>>
>> Jim
>>
>>
> You are right I tried it now, it gives the message (while I do not
> understand the reason, because it makes no sense).
> So the best way is to do it from a live CD.
Yep, that was my thought as well, if it’s the root partition. That’s what
I’ve done in the past.
Unfortunately, I don’t have a live CD on me now. Will the above instructions still work at the (repair filesystem#) prompt? I got a message that the file system on sda 7 was somehow corrupted again. This time I attempted to use the expert repair feature on the 11.4 DVD. At the “update, installation” page I aborted and was offered the expert options, but I had not idea what to choose to repair sda 7. I tried “verify installation” which took about 15 minutes, but it claimed everything was fine.
Then I logged in through failsafe mode again. I got a bunch of questions asking me if I want to fix a long laundry list of items. I said yes to everything which took about 10 minutes. Then it allowed me back into the desktop. All my desktop settings were gone except for a few.
The real problem I’m having now is that I can’t get on the web using Google WiFi. Before this incident, I logged on with no trouble. Now, neither Konquerer nor Firefox can do it. Sadly, I’m composing this message through WinXP. I need to get in through OSS because that’s where I get all my work done.
What settings am I missing in OSS? I checked Konquerer, Network Manager. What files should I be checking for errors? I know that I have to replace Audacity because something has disabled it’s proper operation. I figure there must be other errors I should be checking. Login message through failsafe mode claims that the file systems are clean.
Has it occurred to you that you may have a HDD in a degrading condition? All this is not happening for the fun of some developper, it’s happening for a reason.
Ok what is one sda7? If it is not root then just umount it and run fsck on it. But if it is root you must run from another source. It need not be OpenSUSE Linux any Linux based runnable CD that supports ext4 should be fine.
If the file system is messed up all sorts of thing will be broken. Best not to run that system boot from a CD and fix it.
I installed my Western Digital HDD boot drive October 2010. I don’t see how it could be developing problems already. Seems to me it would be far more than sda7 giving me problems if that was the case.
On 2012-03-20 22:36, HealingMindNOS wrote:
>
> Unfortunately, I don’t have a live CD on me now. Will the above
> instructions still work at the (repair filesystem#) prompt?
From the DVD? Yes.
> I got a
> message that the file system on sda 7 was somehow corrupted again. This
> time I attempted to use the expert repair feature on the 11.4 DVD. At
> the “update, installation” page I aborted and was offered the expert
> options, but I had not idea what to choose to repair sda 7. I tried
> “verify installation” which took about 15 minutes, but it claimed
> everything was fine.
I don’t remember what that does, but I think nothing relevant for your
case, it is an rpm check if I remember right. The automatic repair feature
you may have read about was removed time ago for lack of a maintainer.
You need to run fsck on sda7.
> Then I logged in through failsafe mode again. I got a bunch of
> questions asking me if I want to fix a long laundry list of items.
What questions?
> I
> said yes to everything which took about 10 minutes. Then it allowed me
> back into the desktop. All my desktop settings were gone except for a
> few.
>
> The real problem I’m having now is that I can’t get on the web using
> Google WiFi.
Your real problem is that you have disk corruption, the rest is menial. I
would download ASAP the disk check utility from your disk manufacturer,
like seatools from Seagate, and thoroughly check that disk.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)
On 2012-03-20 23:06, HealingMindNOS wrote:
>
> gogalthorp;2449937 Wrote:
>> Ok what is one sda7? If it is not root then just umount it and run fsck
>> on it. But if it is root you must run from another source. It need not
>> be OpenSUSE Linux any Linux based runnable CD that supports ext4 should
>> be fine.
>>
>> If the file system is messed up all sorts of thing will be broken. Best
>> not to run that system boot from a CD and fix it.
>
> sda7 is my home dir. sda6 is my root dir. Do I perform the above
> commands in a terminal as root? i.e.:
Yep. Close your graphical session, switch to text mode (ctrl-alt-f1), log
in there as root (yes, in text mode it is appropriate), and do those commands.
You say your disk was bought in 2010. What is important is the number of
hours of use - but you have to know that devices fail most at very early
age and at old age. They call it the bath-tube curve. Good manufacturers do
a burning test to their devices to crack those early failures when they are
still in the factory, but some do get through and fail when used at the
customer.
And yes, it can affect a partition only. It may be surface failures
affecting a region of the disk.
Do the tests I told you in my previous message.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)
I just RMA’d a 1 TB Western Digital drive that I installed last May. Only one partition had problems, but it was definitely a hardware issue–both smartctl and WD’s Data Lifeguard reported read failures. You really should check your drive with one or the other (or both).
Unfortunately, the extended test using same on my WD HDD has reported bad sectors - and that I will lose all data in those sectors upon repairing them. I am ****ed if I do, and ****ed if I don’t. I know when this happened too - the day of that solar storm. MY PC suddenly went off although I have a UPS protecting it. I hate moving, but it looks like I have no choice. WD’s Data Lifeguard “may” be able to repair it, but I’ll have to start from scratch anyway.
I’d like to thank everyone of you for all of your help.
I just found your discussion here and I would like to mention that the openSUSE initrd supports a switch “shell=1”. If you add this to your kernel command-line (in grub), the initrd will spawn a shell just before pivoting the root filesystem and calling /sbin/init. There you can unmount /root and then call fsck. The fsck binary for the root fs should be included in the initrd by default. Working inside this initrd environment is a bit bumpy (I think e.g. PATH is not properly set), but you can solve lots of problems much quicker than burning a livecd, e.g. a broken root fs, or a lost root account password.
Maybe this helps others who stumble over this thread due to similar problem. – Yarny
> Unfortunately, the extended test using same on my WD HDD has reported
> bad sectors - and that I will lose all data in those sectors upon
> repairing them. I am ****ed if I do, and ****ed if I don’t. I know when
> this happened too - the day of that solar storm. MY PC suddenly went off
> although I have a UPS protecting it. I hate moving, but it looks like I
> have no choice. WD’s Data Lifeguard “may” be able to repair it, but I’ll
> have to start from scratch anyway.
The data is probably lost already, but you don’t know yet :-}
You can try, if you have the space, to copy the files to another disk. If
you get read errors and the file is important, try dd_rescue.
> I’d like to thank everyone of you for all of your help.
Welcome.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)