using SLES 11.3, I updated from 10.4.
Now when the system boots it spits out an error saying either cannot read system clock or time is wrong, and the time in linux sets to 00:00:00 and the year goes to 2013 or 2005... sometime in the past.

When the system shuts down, it tries to write the time in linux back to the hardware system clock, and it does that wrong too then screwing up the hardware system clock.
I found in system services I can turn off the boot.clock service so it does not write and screw up the system clock.

so right now my hardware system clock stays correct and does not get modified. But how can i fix linux from either incorrectly reading the hardware system clock, or prevent it from reading it in the first place?

Set up NTP and pull time from a regional pool.
At least, that's what I'd try first.

can you elaborate on "regional pool" and how to set up ntp ?
I've never had luck with ntp because i'm behind a firewall and proxy server.
but i have other linux systems running sles 11.3 that run fine and hold correct time, and are on the same network switch so if i can't pull time from like time.gov or whatever, how would i get it to pull from my other linux system?

Sure.
Popular NTP servers can be a pain in the neck for their owners because of the traffic they generate. In addition, there is always some malicious nitwit who wants to create problems for others, so they can end up getting DOS blasted by the script kiddy crowd.
NTP.org got around that by creating "pools". IP addresses that are forwarded to one of a group of servers based on load balancing.
Much nicer than incessantly pestering the guy who runs the atomic clock in Boulder (I'm sure he has enough to do already).

You need to invoke ntpd (it's a service, so it's persistent so in the *nix world that makes it a daemon) with the '-g' switch which precludes it from looking at system time for comparison (because it will freek out if it sees to much "drift" between the system time and the ntpq response).
I'm not the guy to ask about configuring NTP in Susse, however try "man 'ntp'/'ntpd" or "man -k/K 'ntp'/'ntpd') and that will direct you to the appropriate config files in Susse. Apparently "yast" has an NTP configuration capability (I always configure network services by hand because you're always better off understanding the technology rather than understanding the tool however if yast can make it happen, who am I to judge? )

EDIT: Sorry I didn't address this at first, my mistake.
You'd have to set it up as a local NTP server, which isn't hard either although I'm afraid if I gave you directions it would just confuse you. I'm not a SUSSE guy (though I do like the little lizard). I'd suggest you make the proxy the network clock or a router that can bypass the proxy and still reach your servers.

playing with ntp stuff now.
suse 11 ntp : https://www.suse.com/documentation/s...netz_xntp.html

can anyone can explain things in 50 words or less?
I don't feel like scouring documentation

here's a quick shot of what happens during boot up

Second 0 (sorry, second 3)
The beginning of the Unix Epoc.
You call it an error but really.... it's almost magical


Be proud, my son.....

It sounds like the CMOS battery is dead, and that's why the time is getting reset every time the computer is rebooted. Are you able to open the computer up and replace it? On most computers, it's a watch battery somewhere on the motherboard.

it's a big rack server. before linux even loads, i'm in on the system controller and that holds the date & time fine.
so it's not the battery,
and when i ran sles 10.4 it reads the hardware clock fine.
so pretty sure something changed going into sles 11, and the machine is old having a deprecated time syntax or something,
that's my best guess. i need to scrounge another hard drive, load sles 10.4 onto it, and verify.