Reinstalling Windows

Windows 7 laptop with Windows product key sticker on the bottom.

Hard drive failure, so cannot use recovery partition.

Model: ASUS K43U – old enough that Win 7 would be a better fit for it, but little choice AFAIK. (Well, I could install Linux etc, but I have a Linux laptop. This is meant to be my Windows machine, for working with the non-geek world.)

  1. Went to a working Windows machine
  2. Went to https://www.microsoft.com/en-us/software-download/windows10 and clicked ‘Download tool now’
  3. Scrolled down to ‘Using the tool to create installation media’
  4. Put a working 8GB USB stick into the computer
  5. Ran the tool (needs admin)
  6. Got 2 options (1) update this computer (2) make installation media
  7. Chose (2)
  8. Then can do an ISO or USB. Chose USB, chose the relevant drive (E:) and left other things as defaults. (One might need to change to 32-bit Windows 10 for the download, I guess, if that suited the hardware.)
  9. Waited for a while as it downloaded and wrote to the USB and stuff
  10. ‘Your USB flash drive is ready’ – clicked next, and safely ejected it, put it into the laptop. Put the laptop on an Ethernet connection
  11. Booted laptop, holding down F2
  12. Enabled UEFI, then saved, exited, rebooted, reentered BIOS and it found the USB stick and added it to the boot list automagically; edited the list to put the USB at the top, saved and exited
  13. Booted into install program!
  14. Set language and hit Go (‘Install now’)
  15. Provided product key off the sticker on the bottom
  16. Accepted licence
  17. Chose: ‘Custom: install Windows only’ (this is not an upgrade, so did not choose the Upgrade option)
  18. Partitioning GUI – used the whole disk; deleted any existing, created a single new one, agreed that Windows might need to create other partitions
  19. Clicked ‘Next’, and away it went, copying stuff and so on
  20. Click through some menus, making selections (eg keymap)
  21. It rebooted, still with USB stick in …?
  22. But all is well. Continued with the installation
  23. Chose various privacy options (mostly turned things off)
  24. No thanks to Office trial
  25. Deleted a bunch of programs I don’t want
  26. Updated and kept removing cruft
  27. Now as usable as Windows will get on this old machine

OK, that’s not a bad process. It is good of Microsoft to make recovery as easy as that. And they still provide Win10 to users of Win7; not that I like Win10, but it’s nice to be able to install something.

 

Good on ’em. 

Shutting down cleanly — avoiding ‘recovering journal’ messages

The problem below went away with the upgrade to the Debian 9 series.


Even after I shut down using the proper route, like via the GUI menu or running

$ shutdown -h now

when I restart the machine I get a message saying something like:

Fri Aug 2 13:27:51 2016: [....] Checking file systems...fsck from util-linux 2.20.1
Fri Aug 2 13:27:52 2016: /dev/sda6: recovering journal
Fri Aug 2 13:28:03 2016: /dev/sda6: clean, 279129/52994048 files, 48724910/211963648 blocks

In other words, the journal does not match the file system state, and something was not properly put away and Debian feels compelled to recover the journal. Sometimes the recovery takes minutes rather than seconds, and I get messages about ‘clearing orphaned inode’. I have a pretty vanilla install of Debian (old stable) and I don’t see why this should be the case. I should add it’s been this way for years and does not really matter, but I thought I’d see if I could fix it.

Took a look at:

http://unix.stackexchange.com/questions/53252/how-to-read-the-messages-during-power-off

Went to /etc/init.d

$ cd /etc/init.d

Edited two scripts

$ sudoedit halt
$ sudoedit reboot

Added ‘sleep 10s‘ just before the halt/reboot line

in halt it now reads

        log_action_msg "Will now halt"
        sleep 10s
        halt -d -f $netdown $poweroff $hddown
}

NOTE: Tried removing ‘-f’ as well.

and in reboot (which I am less likely to walk away from)

do_stop () {
        # Message should end with a newline since kFreeBSD may
        # print more stuff (see #323749)
        log_action_msg "Will now restart"
        read -p "Press enter to reboot" reply
        reboot -d -f -i
}

Both give a chance to read the log messages, and give the OS a few seconds to finish cleaning up after itself. Shut down always seemed to happen very quickly, and sometimes there would be little red error messages during shut down suggesting that file systems had not been properly unmounted, but instead of waiting or retrying the system would halt anyway. I don’t know much, but that did seem to correlate with getting the journal recovery message on the following boot, as if the computer was shutting down before it was finishing tidying up. Hence the idea of a 10 second wait or pausing for input during halt/reboot.

Seems to have reduced but not eliminated issues. I think things need to change within the halt commands.

No recovery.