sqlite error: disk I/O error

[Scott Haneda]

On Jul 9, 2010, at 7:11 AM, Daniel J. Luke wrote:

> On Jul 9, 2010, at 1:53 AM, Scott Haneda wrote:
>> Going a bit off topic here, but S.M.A.R.T. is not all that smart, and there could be data that is secretly failing and becoming corrupt, yet passing all tests.
> 
> I've had good luck with smartmontools (available from MacPorts ;-) ).
> 
> Adding something like:
> 
> DEVICESCAN -m my_email at my_domain -s (S/../.././02|L/../../6/03) -M exec /usr/bin/mail
> 
> to smartd.conf
> 
> So far, every failing disk has generated an error email some time before it became totally unreadable (which makes it less likely that I'll have to restore from backup, so usually saves me some time). Nothing is going to be able to give you 100% warning, though...

This appears to be a significantly complex set of tools.  I read over the man pages and the conf file. There is an initial test you run `sudo smartctl -t short /dev/disk0s2` which when run, from what I can gather, sets up the device in a way that prepares it for automated running in the future.

Instructions say to keep running `smartctl -t short /dev/disk0s2` until you get results, which I have been trying:
  smartctl 5.39.1 2010-01-28 r3054 [x86_64-apple-darwin10.4.0] (local build)
  Copyright (C) 2002-10 by Bruce Allen, http://smartmontools.sourceforge.net

  === START OF OFFLINE IMMEDIATE AND SELF-TEST SECTION ===
  Sending command: "Execute SMART Short self-test routine immediately in off-line mode".
  Drive command "Execute SMART Short self-test routine immediately in off-line mode" successful.
  Testing has begun.
  Please wait 2 minutes for test to complete.
  Test will complete after Fri Jul  9 15:17:34 2010

Each time I run it, that time to completion moves forward, I don't think it will ever finish.  For example, here is a second run:

  Testing has begun.
  Please wait 2 minutes for test to complete.
  Test will complete after Fri Jul  9 15:18:37 2010

And a third...

  Testing has begun.
  Please wait 2 minutes for test to complete.
  Test will complete after Fri Jul  9 15:18:46 2010

The time is pushed forward, is this correct?  I was under the impression it should eventually finish.  `ps` has nothing running.  And then there is the much longer test to set up the environment and get it all ready to run based on the conf file.

How did you arrive at your conf file options?  Out of curiosity, what does your email report look like?  Do you notice any system performance degradation when the scan is running?  And all automation is handled internally by the dameon, there is no launchctl or any other schedulers that I need to look into to make sure that it keeps on running?

I want to make sure the smartd is running, but can't seem to find out what keeps it running all the time, past reboot, on schedule.  Thanks for any pointers.  If this gets too out there for a MP post, shove me away to the smartmontools mailing list, and I can take it up there.  I had no idea of this tool, and since you have first hand experience with it, I'm hoping we can deviate a little here to get on the same page.

Thanks Daniel, looks like a really valuable tool to have.

-- 
Scott (* For off-list contact, replace talklists@ with scott@ *)