Forum: LinuxTotal Replies: 12
Author Content

Jan 24, 2011
6:14 PM EDT
When I say everyone in my family runs Linux, I'm not kidding. Even my dad, whose only use for a computer is an easy game of Solitaire. I configured it to bypass all logins on boot, going straight to a GNOME desktop and Solitaire. It isn't connected to any network, so I don't worry about updates or security. The only personal information on it is that Dad's preferred Solitaire game is Klondike.

As of today, the uptime on his system is 394 days.


Jan 24, 2011
6:15 PM EDT
I just rebooted my machine yesterday to change out a hard drive. :(

Jan 24, 2011
6:17 PM EDT
gus install pysolFC, it's in most repositories. Klondike is but one of the solitaire games available in pysolFC.

Jan 24, 2011
6:39 PM EDT
The server I put online in 1995 looped the old ~400 uptime counter ... twice.

It was rebooted exactly thrice between 1995 and 2003 when it was decommissioned.

To move it, to upgrade to the 2.0 kernel, and California brown-outs got it once.


Jan 24, 2011
6:45 PM EDT
I heard once of a system that got moved, but not powered down. The UPS was part of the building, and so couldn't move with the system. So they brought the system down to single-user, fastened the batteries directly to the case with duct (gaffer's) tape, then opened up the AT power supply and clipped the batteries directly to the load-side.

It kept the system going long enough for the transport, no power-down necessary. I guess their uptime was important somehow.

Jan 24, 2011
7:24 PM EDT
@gus - I completely understand. Old time unix folk know that uptime is one of the things that separate unix class OSes from peecee class OSes.

How's this for uptime?

obsidian: /home/jjs 
(tty/dev/pts/1): bash: 1002 > uptime
  3:22pm  up 1618 days 18:22,  2 users,  load average: 0.93, 0.96, 0.94
obsidian: /home/jjs 
(tty/dev/pts/1): bash: 1003 > uname -a
Linux obsidian 2.6.5-7.267-bigsmp #1 SMP Wed Jun 21 10:50:51 UTC 2006 i686 i686 i386 GNU/Linux
obsidian: /home/jjs 
(tty/dev/pts/1): bash: 1004 > 

Jan 24, 2011
11:18 PM EDT
Shameless watt-wasters!

$ uptime 19:18:09 up 54 min, 2 users, load average: 0.00, 0.03, 0.00

Jan 25, 2011
1:38 AM EDT
32 days for my home desktop 50 days for my home server (kernel upgrade) 112 days for my remote server (also a kernel upgrade)

My laptop and work machines are shut down when I'm done, so it's rarely more than a day.

Jan 25, 2011
6:37 PM EDT
For me, my home server only goes down for kernel updates, same with the HTPCs on the network (except one that gets rebooted every time it runs Skype). Touchscreen netbook ~12 hrs at a time, while at work (but runs the same chipset as one of the HTPCs). I won't count the routers, even though one is a Micasaverde Vera, & just gets used for that (its been rebooted 2-3x in as many years)...


Jan 25, 2011
7:12 PM EDT
Thanks to ksplice uptrack, I can now apply linux kernel updates without rebooting - so my next goal will be for that elusive 2000 day uptime..

Jan 26, 2011
4:59 PM EDT
Uptime is sóóo 20th century, I finally configured TuxOnIce2 a few months ago and very happy with the results! Why leave in standby if it wants to hibernate?

Jan 26, 2011
5:09 PM EDT
As far as my Linux desktops go, I power down every day. I want a clean system when I'm working. Firefox especially gets pretty gummy after a day. I'm sure that quitting and restarting the app helps, but a full reboot can't hurt.

I keep my Windows XP box running all the time, but I still reboot it every day. That's a system that benefits from frequent rebooting.

I'm not personally doing the admin work on any servers, but I imagine that cron jobs can keep them running well w/o rebooting. I'd still be inclined to reboot after a kernel update. The downtime is only a few minutes per update.

Jan 26, 2011
6:44 PM EDT
I was talking about servers, with my long uptime example, and 2000 day goal - naturally the desktop machines get booted a lot more, especially given all the experimentation we do with bleeding edge kernels.

But on the server side, the whole point of ksplice is that the reboot is no longer needed as the new kernel is spliced in while running. Cool stuff. I've been using it and have gotten several kernel updates, no reboot and everything is solid, as I'd expect it to be.

When I've got a linux machine hosting a dozen important production VEs, the last thing I want to do is reboot.

You cannot post until you login.