system becomes read-only

Portal Home > Knowledgebase > Articles Database > system becomes read-only

Posted by hbhb, 01-01-2010, 01:29 PM
Out of sudden today my server turns into read-only mode. I booted as root and run e2fsck /dev/sda2 and that fix the problem Few hours later, it becomes read only again, and I had to reboot to run e2fsck again. When I check my screen it shows: BUG soft lockup - CPU #1 Stuck for 10s! [quotaoff:5754] etc etc... What am i dealing with here?
Posted by PCS-Chris, 01-01-2010, 01:31 PM
If your filesystem is going read only then its going to be a dojy disk, raid controller or motherboard in most cases. Could be memory as well though.
Posted by hbhb, 01-01-2010, 01:45 PM
no exact way to find out?
Posted by PCS-Chris, 01-01-2010, 02:02 PM
Boot into a rescue CD and run some checks.
Posted by UNIXy, 01-01-2010, 02:20 PM
This is most likely a disk issue (from the controller to the disk including environmental factors like heat). Especially that it stuck at quotaoff, a disk utility. Is /dev/sda2 attached to the system or a remote LUN? If it's a LUN (like SAN or iSCSI) check the paths. What does your /var/log/dmesg say? Regards
Posted by VIPoint, 01-01-2010, 02:22 PM
Looks like a faulty disk... Try to get the sever up and take backup of the data ASAP
Posted by jalapeno55, 01-01-2010, 03:16 PM
I second this, once, it could be something else, twice, and the disk is probably on it is final days, copy the accounts to a new server, or run a fresh backup (if you do backups to another drive or alternate location) and have your datacenter restore ASAP.
Posted by jai_hoo, 01-02-2010, 12:50 PM
As this is a recurring issue.. server seems to be having a dieing hard drive... so the permanent fix would be to take the current backup of your server and migrate the accounts to a new disk....
Posted by aeris, 01-02-2010, 01:50 PM
I recommend running a long SMART test on the disk. That will identify quite a few (although not all) problems. You can start it by using: smartctl -t long /dev/sda Then about three hours later when it's done: smartctl -l selftest /dev/sda If it fails, you most definitely will want to replace the disk ASAP. Note that it works the disk fairly hard, so if your server is loaded, it will slow things down. You should also make sure that your backups are updated *before* you run the test
Posted by nitaish, 01-03-2010, 09:37 AM
It may also be a bad drive. My recommendation would be to get the data backed up and hard drives replaced.
Posted by ianeeshps, 01-04-2010, 02:46 AM
if ur getting disk read only problem then it is high risk to continue with this. There is a change to loose full data. reload is with a new hardisk and mount the current one as another and resync the data..ASAP

Add to Favourites Print this Article