System Disk Failure

From Ed's Mediawiki
Revision as of 21:12, 1 January 2013 by Eburdick (talk | contribs) (Recovery Log)

Update 12/31/2012:

I got some errors reported on drive C: from the HP hardware diagnostics. Just in case this is a real problem, I decided to make a new disk image. The backup failed with this message:

The operation failed due to a device error encountered with either the source or the destination. If
the source or destination volume is on a disk, run CHKDSK /R on teh source or destination volume,
and then retry the operation. (0x8078012D)

Additional Information:
The request could not be performed because of an I/O device error. (0x8007045D)

A web search for the I/O device error code yielded an suggestion by a microsoft person to turn off "shadow copies" on the drives. The forum is here. A search for 0x8078012D brings me to this forum, which also has some interesting suggestions.

Now running CHKDSK /R on drive U: Drive U: checks out fine.

Now checking out drive C: which needs to be done during boot...bad news on this. The drive is in worse shape than I assumed. CHKDSK fails in progress with a blue screen crash. The blue screen just flashes up for less than a second, but I managed to take a picture of it. The error is BAD_SYSTEM_CONFIG_INFO, which web research shows is a corrupted registry. After the blue screen, it tries to reboot, which results in a bad mbr error, so this thing is probably toast.

Backup policy change

I have been telling people that I periodically make an image of the system drive, but I have not been making them often enough. To top that off, I deleted the most recent image in order to make a new image of this drive before replacing it. Of course, this failed, so I no longer have an image. Dumb move. It is a good time to do a clean build anyway, but I need to start doing image dumps once a week or so and always make sure the most recent one is available.

Replacing the drive

Having not such good luck with the Seagate drives, I am trying a Western Digital Caviar Black drive, which get good reviews, and have a 5 year warranty.

Model: WD1002FAEX
Serial Number: WCATRA155218

This is a 1TB 7200 RPM SATA drive with a dual processor. According to overclocking and gaming sites, this has higher quality build and is significantly faster than moderately lower end drives.

Recovery Log

  • Physically installed new drive
  • Booted with recovery disk 1
  • Ran diagnostics. Everything passed except the bad drive. The new drive and the V/U drive passed tests.
  • Disconnected everything except keyboard and mouse, plus new drive. Temporarily disconnected other internal hard drives.
  • Started recovery. It ran without any problems. Did some initial network stuff, but need to get rid of the home group stuff.
  • Downloaded Firefox and created Kits directory tree. Now editing this file with the recovered machine.
  • Disable homegroups because I only have one other win7 computer.
    • Right click on "Homegroup" in the explorer navigation pane.
    • Select "Change HomeGroup settings" and click on "Leave the homegroup..." in the resulting dialog. I did not have to do this because I never joined it.
    • Go to Services and change the Startup type of "HomeGroup Listener" and "HomeGroup Provider" to "Disabled"
  • Set up SMB network workgroup
    • Right click properties on Computer in the start menu
    • Click Change settings on the dialog and fill in blanks.
      • Computer description: Ed's Desktop
      • Computer Name: ejbdesk1
      • Workgroup: AMR
    • Reboot - The reboot really went fast. I wonder why.
  • Set IP address to a fixed 192.168.1.8
    • Click Local Area Connection in Network and Sharing Center
    • Click Properties in the resulting status dialog
    • Select Internet Protocol Version 4 (TCP/IPv4)
    • fill in the blanks. The router is at 192.168.1.1, so that is the default gateway.
    • For the DNS server values, I copied the values from the WAN connection on the router.
  • Updates have been coming in the meantime, so I am prompted for a restart. Restarting...
  • Time to reconnect the other internal hard drive. This is a 2TB drive that hosts V: and U:
    • Power down, reconnect drive, power up and adjust setup to make sure the right disk is being use for booting.
    • Reassign disk letters to U: and V:
  • Change power settings to keep monitors from powering down automatically. This is to avoid having the machine freeze up because of Windows bugs.
  • Restore C: drive files
    • Download and install CrashPlan
    • Plug in drive X. While I was at it I plugged in all of the USB cables.
    • Set up drive letters.
    • Start up CrashPlan and log in.
      • CrashPlan noticed that it did not recognize this computer and gave me an option to adopt it in place of the old EJBDESK1. Big improvement over what happened last time. Adoption was successful and synchronization has started.
    • Ready to restore after resync
  • Set up user ejbtech as administrator with password. Set up ownership and permissions on private directories.
  • Norton Internet Security 2010 was already installed, so I brought it up and run live update. A window popped up to show the serial number of the installation. I am assuming this is how the system knows who I am. The update started with the knowledge of my current subscription.
    • Logged into my Norton account and did an update check. This triggered a download of the latest version, which installed and asked for a restart. Restarting...
    • After restart, Norton Internet Security upgrade completed. Another live update did some additional stuff, and now things are stable. Check for updates indicates everything is up to date with subscription good for 217 days.
  • Install Java latest version to support some web pages.
  • Update Silverlite for some microsoft pages.
  • Update Windows Essentials to get Moviemaker. Not really interested in the other stuff (messenger, mail, gallery,) but it comes with it. Need to restart for this...