Best Linux Server uptime?

[thecoolnessrune]

My new employer is 100% cloud. Systems are no longer treated like family, but rather livestock. I would be shocked to see a machine live 30 days let alone have an uptime that long.

Same for one of our clients. They treat systems / applications the proper way since they're newer. I.E., servers (and applications that run on servers) must be treated as if they'll go down at any moment, not like the old school thinking of them running forever. They patch every server, every month. They consistently have the fewest issues of any of our clients, and little to no security vulnerabilities to manage.

My home environment is patched in its entirety every month. If something that I'm running has issues with a certain patch level, then I'll give it a little time to be resolved otherwise I'm chucking the software.

I can definitely see where exceedingly long uptimes (thousands of days) is pretty impressive, otherwise I see it as fairly equivalent to flaunting one's A+ Certification from 1995. Archaic processes from archaic thinking.

 

[dave_the_nerd]

My new employer is 100% cloud. Systems are no longer treated like family, but rather livestock. I would be shocked to see a machine live 30 days let alone have an uptime that long.

That would be nice. Would be nice if anybody around here was willing to either 1) engineer something that would do that, or 2) pay the extra money to buy the HA/Clustered variations of the tools we use.

*sigh*

 

[wirelessenabled]

Your servers are old enough to drink?

That HP-UX box is either old enough to drink or is soon approaching that day.

 

[dave_the_nerd]

That HP-UX box is either old enough to drink or is soon approaching that day.

Christ. I need a drink.

 

[Red Squirrel]

The issue with cloud is when it goes down there is absolutely nothing you can do about it. You are at the mercy of the provider to get stuff running again. They have lot of redundancy so the odds are in your favor, but still, I personally could still not trust it, I just like being in control of my infrastructure. There is also the privacy issue.

 

[Paperlantern]

I can definitely see where exceedingly long uptimes (thousands of days) is pretty impressive, otherwise I see it as fairly equivalent to flaunting one's A+ Certification from 1995. Archaic processes from archaic thinking.

This. Yes. With vulnerabilities being found and patched at an almost DAILY rate, uptime for any server over a few weeks could actually be considered irresponsible.

 

[Red Squirrel]

This. Yes. With vulnerabilities being found and patched at an almost DAILY rate, uptime for any server over a few weeks could actually be considered irresponsible.

Most of these servers with super long uptimes are probably not public facing. Mine arn't, except for my web server and when I do updates I restart all the services (apache etc). With Linux you don't need to reboot because the mouse moved, like in Windows.

 

[sourceninja]

...

 

[dave_the_nerd]

Somebody finally did a hardware audit that involved physically going to all the racks, and they found this thing.

$Last login: Tue Jun  6 08:52:08 2017 from delhi.beer.town
Sun Microsystems Inc.   SunOS 5.10      s10_72  December 2004
You have new mail.
root@crazyhorse#uptime
  2:38pm  up 1360 day(s), 10:12,  1 user,  load average: 0.02, 0.01, 0.01
root@crazyhorse#uname -a
SunOS crazyhorse 5.10 s10_72 sun4u sparc SUNW,UltraAX-i2

 

[Red Squirrel]

Haha nice. My oldest home server is still chugging along. Not my uptime record but getting there. I don't really recall my exact record but was around 1,300 for an online web server. It went down due to data centre UPS failure.

[root@borg ~]# uptime
 20:34:03 up 1096 days,  3:01,  1 user,  load average: 0.00, 0.00, 0.00
[root@borg ~]# 
[root@borg ~]# 
[root@borg ~]# uname -a
Linux borg.loc 2.6.27.25-78.2.56.fc9.x86_64 #1 SMP Thu Jun 18 12:24:37 EDT 2009 x86_64 x86_64 x86_64 GNU/Linux
[root@borg ~]#

 

[adorid]

Oldest server in company where i work. This server is so old with so many bad sectors that restart would kill it so soon we gonna change it.
http://imgur.com/S7thpD2


FreeBSD 5.2.1-RELEASE #2: Wed May 25 22:23:50 EEST 2005

Welcome to FreeBSD!


$ uptime
4:28PM up 4159 days, 11:41, 2 users, load averages: 0.14, 0.06, 0.01
$ uname -a
FreeBSD 5.2.1-RELEASE FreeBSD 5.2.1-RELEASE #2: Wed May 25 22:23:50 EEST 2005

 

[Red Squirrel]

lol nice! Yeah, you don't want to touch it at this point.

 

[Red Squirrel]

Still going strong!

[root@borg ~]# uptime
 06:13:13 up 1144 days, 12:40,  1 user,  load average: 0.00, 0.00, 0.00
[root@borg ~]#

I need to get off my butt and finish migrating stuff off that server though. it's only doing email at this point, but that is a pain in the ass to migrate/setup so I keep procrastinating. Once I can migrate stuff off it I can shut it down and probably save myself about 100w or so.

 

[you2]

With an update of 1144 days how to in you install security patches (esp for hte kernel which must be quite old) ?

 

[Red Squirrel]

Lol the OS on there is Fedora Core 9, there are no patches since it's been out of support for probably close to a decade. It's an internal server so is not really at risk.

If it was a supported OS then there would be patches, the nice thing with Linux is not having to reboot for every little patch. I just restart the services like apache etc. Anything internet facing is a potential risk if it's unpatched.

 

[you2]

Yea but these days I tend to get a kernel update once every 6 weeks. I can skip the reboot for simple application updates but usually run into problems when kernel and system libraries are updated. Hum. Been at least 4 months since the last ssl/ssh exploit. We should be getting one of those soon but at least they don't require a reboot.

 

[mv2devnull]

The "best" uptime is on those servers that you can reboot whenever needed without disturbing services.

 

[TeknoBug]

My dual Pentium Pro server ran from 1998 to 2010 and its longest uptime was just shy of 4 years, I forget the exact numbers but it was a rock solid server with its 6x SCSI array and regularly ran at 100-120C temps.