[geeks] operating systems to replace Solaris

[mail at catsnest.co.uk]

On Mon, Apr 11, 2011 at 3:45 PM, Phil Stracchino <alaric at metrocast.net>
wrote:
> On 04/11/11 10:24, der Mouse wrote:
>>> I can get a two week long run on the cluster running Solaris and not
>>> get a crash which cann't be said for linux or BSD
>>
>> I wonder what I'm doing right - or you're doing wrong - then. B Sampling
>> the uptime of my most heavily used machines, all running NetBSD, I find
>>
>> 34 days
>> 41 days
>> 57 days
>> 5 hours
>> 72 days
>> 35 days
>> 41 days
>> 1 day 22:37
>>
>> The "5 hours" one lost power this AM; the "1 day 22:37" one I was
>> working on Saturday AM - I think I needed to install a new kernel or
>> some such. B I haven't dug through logs to find reasons for the most
>> recent reboots of the others.
>
> Same experience here. B My uptime on this machine (running Gentoo Linux)
> is only 21 days now, because I rebooted recently to load a new kernel.
> Prior to that, other than new-kernel reboots .... hmmm .... I had to
> reboot about five months back to swap out the video card. B At one point
> somewhere during that period I had one abnormal reboot forced by a
> software issue, but I don't remember now what it was or exactly how long
> ago.
>

" What I can say is that as a general rule, my uptime is limited by how
 often I update my kernels, not by any stability issues. B "
<rant>
I am probably approaching this from a narrow view point(1) (and also
contributing nothing to the original thread) but this is my problem
with Linux...
Why do we need to do kernel updates for a service that is in place and
has had no major changes?
Don't get me wrong I believe in appropriate security patching (if
those vulnerabilities affect the service in question).
If the service has been designed correctly with security in mind in
the first place, the fact that if someone who is already on your
server as a trusted user could get root by sending some random strings
to a messaging service is nothing as they should never should of got
into that trusted status in the first place.
This is why i have always loved Solaris as even though it can take
more effort to set up once its going there is very little to go wrong.

The change to the license after the Oracle transision has left a lot
of people with very difficult times ahead, Oracle seem to have
forgotten the people that keep Solaris alive and that were driving it
into the community.
</rant>

and ok ssh and dns is its only main job's but...
u5# uptime
 18:39pm  up 1040 days 23:52,  22 users,  load average: 0.44, 0.10, 0.04

1)
I am in general responsible for "long running services", eg
applications and services that stay up with minimal downtime for
years.

Ritchie,
--
Fnord...
<-- http://23.me.uk --> <-- http://wiki.s23.org --> <--
http://twitter.com/23me -->
<- -Only the lavender prairie dogs know the key to unlock the secret
of the tumble-weeds. -->



On the previous
> machine to this one, that was not the case, because even with an uprated
> power supply it was a bit shaky when maxed out of memory; but that was a
> hardware stability issue, not software.
>
>
> --
> B Phil Stracchino, CDK#2 B  B  DoD#299792458 B  B  ICBM: 43.5607, -71.355
> B alaric at caerllewys.net B  alaric at metrocast.net B  phil at co.ordinate.org
> B Renaissance Man, Unix ronin, Perl hacker, SQL wrangler, Free Stater
> B  B  B  B  B  B  B  B  It's not the years, it's the mileage.
> _______________________________________________
> GEEKS: B http://www.sunhelp.org/mailman/listinfo/geeks