Re: [tlug] longest uptime boxes

[Josh Glover <jmglov@example.com>]

Quoth Nguyen Vu Hung (Mon 2002-10-28 02:14:44PM +0900):
> 
> On Mon, 28 Oct 2002, Jim O'Connell wrote:
> 
> > Uptime is a dangerous addiction that is best avoided.
> 
> so why?

I agree with Jim here. When you start making bad decisions to avoid rebooting,
you will see what I mean.

Uptime means nothing to me.

> This is one of the reason why I dislike linux ;):
> When we upgrade kernel, we have to reboot to active new kernel.

How many OSes allow you to upgrade a running kernel? It would have to be a
pretty modular kernel, wouldn't it. And what if you needed to upgrade
something in the tiny core?

Reboot.

This is an open issue in embedded systems, where downtime is really
unacceptable. That is the point of technologies like Sun's HotSpot VM, which
allows a running system to be upgraded. Erlang, a language first developed
at Ericsson and now Open Source, also addresses this in an interesting way.
For more on Erlang:

http://www.erlang.org/white_paper.html

> Of cource, big company like a bank cant reboot their server ;)

Sure they can. When you are running a high-availability site, you have
failover. So, you upgrade the spares in your system, one at a time, and
failover keeps your site up.

> you should upgrade your software first, Jim

Why?

There are two reasons to upgrade, IMO:

a) Bugfixes that you must have (e.g. security-related)
b) New features that you really need

Chasing the newest versions of all the pieces of software that makes up a
modern Linux distro is a waste of time in the best case.

The sysadmin's motto: "if it works, do *not* upgrade it". The exception is
security, of course, but a security issue really means that it does not
"work".


-- 
Josh Glover <jmglov@example.com>

Associate Systems Administrator
INCOGEN, Inc.
http://www.incogen.com/

GPG keyID 0x62386967 (7479 1A7A 46E6 041D 67AE  2546 A867 DBB1 6238 6967)
gpg --keyserver pgp.mit.edu --recv-keys 62386967

Attachment: pgp00106.pgp
Description: PGP signature