Re: [vox] Uptime :) <smirk>

[speck@blkmtnMAPS.org]

Actually yes (depending).

For straight file servers, I have had that happen.  As long as the app you are running on the server doesn't have memory leaks, you don't inadvertantly run out of disk space for a systen file, and your database doesn't corrupt, they will do this.  However bad luck, not having time to pay attention properly will happen.

One disavantage on NT/2000 is more reboots to patch, etc.  However, a full service pack, or hotfix is actually the equivilent of recompiling the kernel, so a reboot we go.  The architectures are different, so there will be trade offs.  

Also, about uptime, 'almost' so what.  I answered the original question with the nearest system I had, all the other servers have been through a HW upgrade more recently.  My home w2k system -dual 300 celeron, file server, Exchange had been up for a year at the time I moved, so it got shut down then.

I know others that have more difficulty than I do, but also several who have the same experiance I do with NT/2000.  A lot depends on the quality of hardware (HCL?), UPS, the initial desicians during setup and the usage.

Also, many companies try to 'cheap' it out in ways that cost them more money in downtime in the longrun.  
Windows admin's through a GUI, must be easy (not!)  They will pay less for certification without experiance (earned in a boot camp) than experiance with or without certification.  The disadvantage to this approach is that many newbies will try straight book (cookie cutter) approaches to things as oppossed to best known practices to setting up systems.  Often, they will not be able to adaquetly prepare for a server roll out because they don't have the experiance to know where to look for answers.  As a result, you get odd/inconsistent stability issues down the road.  Or someone inherits a network with piss poor documentationm, odd quirks and a history of poor performance. 

That still does not mean that Linux servers don't outperform Windows in some areas.  It really just means that uptime is not such a big deal.  With one exception this year, all my downtimes where planned.  Some were for maintaince that did not involve a reboot, some did, it's always a mixed bag.  (ps, I have seen Solaris server die spontanisously, the UNIX admins had a bad week)

'Sides, I still can't get those uptime results with Linux servers.  On one, it was a HW issue, currently, I suspect its an APM configuration issue.  My Mandrake 8.2(?) box just dies on me.  I haven't had time to research it yet.  

At this point I am still better with Windows than Linux, but I have been paid to support Windows for 5 years now.  As my performance directly influenced my pay, that was one serious incentive to get better.  :)

Now, I am working on getting competent with Linux.  After I get competent, than I will work on getting good.  I figure if I learn Linux, then in a few more years I will definatly have more options.  Besides, many Unix admins make more than NT/2000 admins.

Regards,
-sp

On Thu, 20 December 2001, Zach Johnson wrote:

> 
> ok, can NT do...
> 
> [root@line /root]# uptime
>  11:07am  up 202 days, 49 min, 24 users,  load average: 1.21, 1.11, 1.02
> 
> [root@cube /root]# uptime
>  11:07am  up 219 days, 23:39,  5 users,  load average: 1.15, 1.03, 1.01
> 
> [root@helix /root]# uptime
>  11:09am  up 307 days, 14:23,  8 users,  load average: 2.00, 2.00, 2.00
> [root@helix /root]# 
> 
> 
> ? :)
> 
>   Zachariah D. Johnson        Phone : 1-530-752-8801           
>   Systems Administrator        E-Mail: zjohnson@math.ucdavis.edu


_______________________________________________
vox mailing list
vox@lists.lugod.org
http://lists.lugod.org/mailman/listinfo/vox