Getting thrown into read-only mode - not able to remount

archtoad6 · 06-06-2007, 09:14 AM

Be patient. I tend pipe smartctl output 1st to tee to save it to a file & then to less to have a nice scroll-able screen display:

Code:

smartctl -a -t long -d ata /dev/sda  | tee smartctl.log  | less -S

dleangen · 06-07-2007, 01:46 AM

Thanks for the tips.

I never expected to get such great help. I'm impressed by the peeps on this forum!

So anyway, the results of smartctl tell me that there are no problems on my disk.

I also "yum updated" the kernel (yum reported that there was nothing to update).

So, I'm still at a loss as to why my system keeps getting thrown into read-only mode.

Any other leads I could follow up on?

macemoneta · 06-07-2007, 10:27 AM

If the kernel and other software are up to date, then something is writing directly to the partition.

- Are you using VMWare with a partition (even if it's not the one having the problem)? When run in this mode the virtual systems and VMWare have direct access to the drive, and a bug can cause a bad write.
- Can you swap the drive controller card? I've seen the error correction logic fail silently, so bad writes occur with no other indication.
- It can also be a failure on the drive electronics, but if it passes SMART self-test, this is unlikely.
- Have you checked your drive, CPU and case temperatures? An overheating component can do bad things.

dleangen · 06-07-2007, 07:24 PM

Interesting... the one about overheating seems to be the most probable to me.

The problem has occurred before (more than once) during heavy activity. It's possible that this heavy activity (disk, CPU or other) caused the temperature to rise above some threshold and this caused a bad write...

I've never checked for this type of thing before. Is there some utility I can use to monitor and log system temperatures?

If this turns out to be the problem, I guess I'll have to find some kind of cooling device...

macemoneta · 06-07-2007, 07:45 PM

You can check drive temperatures with (for example):

smartctl -A /dev/hda

On my drives, the temperature is attribute 194:

Code:

194 Temperature_Celsius     0x0022   084   076   042    Old_age   Always       -       42

The last value on the line is the current temperature.

If you want to check CPU and/or motherboard temperatures (and your hardware supports it), you need to install and configure the lmsensors utilities. Check your distribution software repositories; it's probably available pre-built.

Physically examine the fans and air ducts on the case and CPU. Vacuum any dust, and make sure all the fans are actually spinning.