My uptime is
Submitted by root
| Less than a day |
   19% | 78 votes |
| Less than a week |
9% | 39 votes |
| Less than a month |
18% | 74 votes |
| Less than six months |
28% | 115 votes |
| Less than a year |
10% | 43 votes |
| Over a year |
12% | 50 votes |
| Total 401 votes |
if it's a workstation in your main room and if you have a girlfriend, it's allways less than a day ;)
[ Parent ]
[ Send Message | View Steve's Scratchpad | View Weblogs ]
True.
My laptop barely makes it past four hours on average - and my backup host tends to be up for a week or two tops.
Still my main machine is up for 47 days right now, and my remote webserver is managing around 100 days on average.
Mostly I reboot machines that have been up for around a year on principle; just to make sure they come back up.
(I had a bad experience once with a Sun box which hadn't been rebooted in 700+ days, and when it was rebooted half the services wouldn't start up - lots of things had been started manually and never added to the init scripts.)
Steve
-- Steve.org.uk
[ Parent ]
I make a point of bouncing servers at least once a month after a difficult evening replacing a sparcstation 2 responsible for the entire company's DHCP infrastructure after the disk wedged on power down. Keeps you on your toes and the complacency down.
[ Parent ]
[ Send Message | View Steve's Scratchpad | View Weblogs ]
Once a month seems a little bit excessive! But it's certainly a good habit to get into.
I usually wait a few months unless it's I've just performed a major upgrade. (A new kernel, new networking setup/VPN tunnel, or something otherwise significant.)
Steve
-- Steve.org.uk
[ Parent ]
The secret is to reboot them just before you start making drastic changes, as well as just after!
I learnt this the hard way, having taken down a database server in a big bank, which processed circa 28,000,000 GBP/hour of banking information, for a disk upgrade, and general tender loving care, only to discover there was no "init.d" script for the database.
Turns out the regular admin (away) knew this, and had documented manual start-up procedures for the database, but it was a bit of a shock to bring the server up, have all the system changes work first time, and them discover the main application hasn't even attempted to start.
As it was all was back before the start of the working day (we had Sunday spare - and two backups - just in case). But the lesson was well and truely learnt, that (especially) on strange servers, reboot at the start and note errors and omissions, so you know what you've broken and what was already broken.
Uptime isn't always a good measure. The only Debian box I administer with any signs of instability is the firewall, and it is in identical hardware (nearly) to one that has been up for the 96 days since we installed it - I think it is netfilter to blame. The "flaky" firewall has been up for 6 days longer than my rock solid desktop, that is mains power for you.
[ Parent ]
desktop i dont rarely get more than a month of uptime, dual booting win + lin can be annoying but battlefield2 is just too good ;)
*space available for advertising*
[ Parent ]
[ Parent ]
Is Windows printing really that unreliable?
Sounds like it is time to change the password ;)
[ Parent ]
[ Parent ]
[ Parent ]
[ Parent ]
[ Parent ]