I've just had the maintenance of an old laptop dumped in my lap. The specific issues with this machine are that it hasn't been rebooted in months, it's using a years-old version of Red Hat, and it has had a number of undocumented tweaks or changes applied so the server can cross compile-code for an obsolete embedded processor. The engineer who finally succeeded in bending the environment to his will has left the company, but I can probably email him specific questions, if I have enough details in the question to jog his memory.

I've got to move this box to another building that is less than a quarter mile away, re-assembled it with some other gear, and test everything. I plan to quickly transfer the laptop plug to an UPS, and hand carry the machine over to the new building so it does not need to be rebooted. At some point, of course, I'm going to have to bite the bullet and get this machine back up after reboot. I'm looking for tips to make that procedure successful.

Pre-move, I plan to log in as root and

  1. Copy off and save the "bash_history" file for all users
  2. Do a "ps-ef >> running-process.txt" to archive what is currently running
  3. Follow the advice of Xerxes in his answer to a related question
  4. Noufal Ibrahim suggested (I think) that I remount the currently mounted drives as read-only and then image them. I've never done that before, but I think that option was added to mount at some point.

Once I finally take down the machine, I'll "dd" image the drive, and store an image locally and in the cloud. I might run this image on virtual hardware in the future.

I'll also try to "suspend" the machine, rather than doing a shutdown. (I don't know if that's an option with this version of this OS.)

I've already moved it without rebooting because I need to reassemble the demo and test it in the new location. If it does not work, I need to fix it. By moving it without rebooting it, I separate the "won't work because you reassembled it wrong" issues from the "won't work because something did not start right after reboot" issues. (I.e. test && move && test && image drive + document state of box && reboot && test.)

migrated from serverfault.com Apr 27 '11 at 10:50

This question came from our site for system and network administrators.

  • I'm currently an unregistered user, so I have to post a comment to log in after my question was migrated here from the forum for system administrators (for some strange unexplained reason, curiously moved without comment...) – utnubu Apr 27 '11 at 11:19
  • @utnubu - if you register both your Super User and Server Fault accounts with the same OpenID they should be associated and you'll regain ownership of the question. – ChrisF Apr 27 '11 at 12:52
  • 2
    I do believe this is a violation if the 'what if X gets hit by a bus rule'. Any specific reason why you can't reboot it? – Journeyman Geek Apr 27 '11 at 13:29
  • 1
    @JourneymanGeek days ago I am going to reboot it, and I want to know what to do before I reboot it to be sure I get it back in the state that it currently is. – utnubu Apr 29 '11 at 18:49

Perhaps you should dd the disk images onto a USB drive or something before you shut it down (by mouting the partitions r/o before you do so).

You never know if the the .bash_history and other things which you're copying are enough. I'd be more comfortable with a complete disk image.


Set up another computer with a working environment before doing anything with this one. Computers are cheap, Fedora is free. If you can't duplicate the environment your company is in trouble right now and you should be working on that problem, not fiddling with the old system.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy