Hello Everybody

We recently rented a dedicated server that runs CentOS 5.0,
We are occasionally experiencing downtime because of CentOS switching the file system into read only mode.
When i login to the server via ssh while the file system is mounted read only, most of the commands result in the following output.
I/O Error, Read-Only file system,
In those circumstance we just restart the server and it would work fine.

So we thought to replace the disk,

Is that correct?
Is there any other reasons for this problem to appear?

If the disk should be the culprit, how shall we proceed to clone it in the new drive?

We have access to KVM/IP,
The drive is in LVM configuration, has only single drive.
500 GB capacity . but 2% only used.
No Live CD option,

Please let me know the best way to proceed in the situation.
I am new to server administration.

OK, if there is an error during boot then the root filesystem will be
mounted read only to prevent damage.Just try to scan your file system as a first step.

Step 1: First try to determine your root partition let say /dev/hda1.
Step 2: Boot the system in to rescue mode using Linux rescue CD or Floppy.
Step 3: Then run the command fsck.ext3 your-partition for e.g fsck.ext3 /dev/hda1
It will take some time to fix if any problem in the filesystem.Then try to reboot the system and observe for the system stability.

I was running stock kernel of CentOS 5.0,
Just updated to 5.1.

I heard that this is a compatibility issue with scsi controller,
Not related to the physical disk,

In the kernel i have updated to has lot of fixes for that.

Let us see how it goes,
Crossing fingers.

BTW the following is the error i got before the filesystem got remounted in read only mode.

kernel: journal commit I/O error

following is the output from the dmesg.

end_request: I/O error, dev sda, sector 773379333
sd 0:0:0:0: SCSI error: return code = 0x00040000

It looped for some time.