Re: [LUG] End of an era..

[Gordon Henderson <gordon+dcglug@xxxxxxxxxx>]

On Sat, 9 Jul 2011, Paul Sutton wrote:

> On 08/07/11 22:49, Dava wrote:
> > Gordon Henderson wrote:
> > > My longest running box:
> > >
> > > # uptime
> > >  22:22:21 up 1374 days,  2:42,  2 users,  load average: 0.00, 0.02, 0.00
> > >
> > > just over 3 years and 8 months. I have to reboot it now to install a new 
> > > kernel to support IPv6. (It's a router)
> > >
> > > Ah well!
> > >
> > > Gordon
> > Thats pretty epic Gordon, good work, there is always something 
> > disheartening about rebooting a box after it has hit such a good uptime.
> is there a website somewhere that this can be put on to,  something like the 
> ubuntu or linux counter, so a few questions on the hardware and os version 
> and uptime see if we can find who has the longest uptime. it could be a 
> popular website and people fight for the top spot.

There used to be a linux uptime project, but I've not looked at it/for it 
for some time now.

Is it that important? Hard to tell. I've been criticised for long uptimes 
in recent years, but I've also been praised for long periods of continuous 
service (so you can't win!) - one theory being that if it's not been 
rebooted, then it's not been patched, etc. However nearly 4 years of 
uptime on a box that has been shifting GB of traffic every day, but 
doesn't itself run services... Well, some ciscos don't always last that 
long - I do have a cisco switch though:

drogonSwitch00 uptime is 5 years, 10 weeks, 4 days, 23 hours, 25 minutes

So maybe some are OK!

And with long uptimes, there's always the wory that something has changed 
which might stop it rebooting - e.g. the BIOS battery has gone flat or 
something. (I've had that happen) or some other upgrade might change the 
boot process, so these days, I tend to reboot my servers in a controlled 
manner, at a schedulled time (round about once a year) so that when/if I 
need to power cycle them for any reason, at least I have a good level of 
confidence that they'll "just work".

That box is one of a pair and it's counterpart was rebooted some time 
back:

$ uptime
 10:33:18 up 984 days, 12:38,  1 user,  load average: 0.00, 0.00, 0.00

I needed to prove the fail-over mechanisms at the time, so power cycled 
it.

However I'll be rebooting that one this weekend too. (Along with a few 
other servers as I move as much as I can to IPv6. I short-sightendly put 
IPv4 only kernels on most of them at the time, and now that I have an IPv6 
allocation and my co-lo is getting ready to enable IPv6, (I'll be their 
test customer!!!) it's a job to be done...

Gordon

--
The Mailing List for the Devon & Cornwall LUG
http://mailman.dclug.org.uk/listinfo/list
FAQ: http://www.dcglug.org.uk/listfaq