So, I’m writing this post in the hopes that someone with more knowledge will read it and hopefully be able to track down the issue.
It started last Friday (February 8th) on my husbands Gateway laptop while he was at work. When he got home on Monday morning, he informed me of it – ‘I think I broke my computer…’ After looking at it for a little while, I realized that his file system had been marked ‘read-only’. A bit of googling informed me that this was usually a mark of a dying hard drive. After backing it up, I decided to try formatting and re-installing to see if that would, by chance, fix the problem. And it did! At least, initially. However, as soon as I updated all the packages and installed a couple of programs (Samba, Chromium), it reoccurred. Which made me suspect, initially that his hard drive as in deed dying or dead.
At approximately this point however, the exact same thing started happening on my laptop (a Lenovo IdeaPad). As a result I posted on ubuntuforums.com (http://ubuntuforums.org/showthread.php?s=19c5720cfb4ab8a14e1f6d7b264e6118&t=2115225) and was directed to a bug report (https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1063354), which matches it. It appears to be a problem related to the kernel, though why it has taken several months to rear its ugly head I couldn’t say. In any case, I did a bit more googling and found a couple of ideas on how to fix it, at least temporarily. By booting into a live image of Ubuntu 12.10 I was able to run sudo dpkg-reconfigure -a – which fixed it on my laptop, mostly. It still goes into read-only mode, semingly at random, but a reboot immediately afterwards has, (thus far) fixed it.
The same can not be said for my husbands’ Gateway. After a third re-install of Ubuntu 12.10 (and giving him instructions to upgrade everything *except* the kernel), it worked fine for approximately 2 days, but has now reverted to read-only filesystem status, when booted. However, when I boot into a live image, the files do not appear to be read-only, and sudo dpkg-reconfigure -a has not worked the same way it did on mine :(. I think I’m going to try a fourth re-install (using 12.04), and hope that it remains functioning…
Any other suggestions are greatly appreciated!!
February 14, 2013 at 1:57 pm
Randomly read-only FS and a reboot fixing it is probably the result of the mount as read-only on error feature. If you look at your /etc/fstab or similar you’ll probably find that your partitions are mounted with errors=remount-ro (on error, remount read-only).
One way to trigger this is if the last write to the partition appears to be “in the future”. That can be the result of badly configured system time, hence why dpkg-reconfigure -a would fix it.
February 14, 2013 at 2:27 pm
I do indeed have an errors=remont-ro in one of my partitions… how would I go about re-configuring system time correctly?
February 14, 2013 at 4:19 pm
sudo ntpdate time.nist.gov us.pool.ntp.org ntp.ubuntu.com north-america.pool.ntp.org pool.ntp.org && sudo hwclock –systohc
The first command, ntpdate, will synchronize your system time from an atomic clock. The second, hwclock, is to set the hardware clock from the current system time.
February 14, 2013 at 2:09 pm
Did you try reading SMART data (using smartctl from command line)? Any modern HDD should support SMART which will give you some diagnostic data on your HDD.
I have a similar issue: My Linux sometimes remounts the SSD in read-only mode (happened ~twice this year) but rebooting helps. If you notice that your HDD is read-only try saving the kernel log (dmesg command) and save it (e.g. with an USB thumb drive). That log may help hunting down the problem.
February 14, 2013 at 2:29 pm
Yes, when I do this I get a series of errors like this one:
60 00 00 00 cf 7e 48 00 11:38:29.670 READ FPDMA QUEUED
February 14, 2013 at 2:38 pm
from SMART or from dmesg?
February 14, 2013 at 2:39 pm
February 14, 2013 at 5:27 pm
Get a new hard drive. It can not be the filesystem, because you created a new one (reformatted, installed ubuntu again, etc.). Also, the disk is failing when you are trying to get access to it from smartctl.
Another alternative is something else broken, but it is less likely.
If you still need to be sure, you can also try to boot with sysrecue (http://www.sysresccd.org), and try to run the tests from there.
February 14, 2013 at 7:16 pm
Thank you. Haven’t had it go into read-only mode yet today, we’ll see how this goes 🙂
February 14, 2013 at 8:42 pm
In cases like this I evaluate what is more important: my data (and my time) or save $100?. The latter still leaves you uncertainty that it can get broken at any time, so either you would lose some data or you will be doing more frequent backups.
However, some people are more paranoid than others with their data.
February 14, 2013 at 8:54 pm
Yeah, I already backed up everything of importance that I didn’t/don’t already have saved elsewhere, so I’m gonna sit and see what happens…
February 15, 2013 at 7:25 am
Have a look at ~/.xsession-errors
Also open a terminal whilst the system is running, or eneable ssh, so that you may be able to check dmesg and system logs after the GUI crashes – these messages can not be written into the log of a disk once it is RO, so you will not see tham after a reboot.
It is not a new issue – eg https://bugs.launchpad.net/ubuntu/+source/linux/+bug/515937 – but perhaps you have some software on both laptops that is triggering a heavy hard disk drive write episode and causing a timeout, above the threshold that the kernel is recognizing as a failure.
February 19, 2013 at 6:56 am
I have that bad bug on two Asus Zenbooks. I’m pretty sure the cause of this bug is not a broken hard drive. But have something to do with a buggy power saving mode (may be caused by a buggy Bios). I came to this conclusion, having this problem only on devices that run in battery mode and it did not occur with Kernel < 3.2. It would be interesting to know whether you can confirm that bug occurring only when your computer runs on battery.
February 19, 2013 at 8:38 am
gunwald – interesting. I haven’t experienced said bug in a couple days since following the instructions given by toobuntu about resetting my clock, but I’ve also been sick and sitting in the same spot so my laptops been plugged in. I’ll experiment with unplugging today. My husbands laptop continued to be buggy and I’ve now installed OpenSUSE 12.2 on his laptop and it has been fine ever since, though I’m not sure how much timee its spent in battery mode vs plugged in either. I’ll check and do some experimenting today 🙂
March 5, 2013 at 8:03 am
Its re-appeared with a vengeance as of yesterday. Its also definitely not only occurring in battery mode – I’ts occurred 3x this morning already, all while plugged in. I know I did a rather large update on Sunday night, and it then ran fine for most of yesterday before locking the drive last night once, and now this morning. FWIW Kevin’s system has run just fine ever since I installed OpenSUSE 12.2 on it, which makes me seriously consider doing the same to mine. I’ll just miss the huge selection of easily installed applications on Ubuntu. :sigh: