SEO

"SCO Announces Winners of Longest Running Server Competition at Annual
Customer Conference

"Eckerd and Zenez Recognized by SCO for Longest Running UNIX Servers Eckerd
runs uninterrupted for a decade on OpenServer, Zenez runs on UNIX for 20
years"





When SCO announced this contest a few months ago, I thought they were
looking for the longest *continuously* running SCO server. With all the
occasional routine maintenance that one has to do on a typical system,
I knew I didn't have a chance. I expected the winner to have a system
with a continuous uptime of 2 or 3 years possibly. (No more due to
the mandatory Y2K upgrades.)

But from the content of this article, it seems that they chose 2 worthy
companies who have simply *used* SCO systems on a continual basis,
with breaks for maintenance and upgrades, for 11 and 20 years,
respectively.

Since my company started with Xenix 2.3.4GT back in 1991, I feel a little
bit
left out. Not that I feel that I should have won, because I know that
there are many people on the ng with years longer experience
using SCO in a single company.

So did the right companies win this contest in your opinion?

Bob

Bob Bailin <72027.3605@compuserve.com> wrote:
>"SCO Announces Winners of Longest Running Server Competition at Annual
>Customer Conference

>"Eckerd and Zenez Recognized by SCO for Longest Running UNIX Servers Eckerd
>runs uninterrupted for a decade on OpenServer, Zenez runs on UNIX for 20
>years"


>When SCO announced this contest a few months ago, I thought they were
>looking for the longest *continuously* running SCO server. With all the
>occasional routine maintenance that one has to do on a typical system,
>I knew I didn't have a chance. I expected the winner to have a system
>with a continuous uptime of 2 or 3 years possibly. (No more due to
>the mandatory Y2K upgrades.)

>But from the content of this article, it seems that they chose 2 worthy
>companies who have simply *used* SCO systems on a continual basis,
>with breaks for maintenance and upgrades, for 11 and 20 years,
>respectively.

>Since my company started with Xenix 2.3.4GT back in 1991, I feel a little
>bit
>left out. Not that I feel that I should have won, because I know that
>there are many people on the ng with years longer experience
>using SCO in a single company.

>So did the right companies win this contest in your opinion?


I read it the same way you did.

Poorly phrased, but too late now.

--
tony@aplawrence.com Unix/Linux/Mac OS X resources: http://aplawrence.com
Get paid for writing about tech: http://aplawrence.com/publish.html

Bob Bailin wrote:
>
> "SCO Announces Winners of Longest Running Server Competition at Annual
> Customer Conference
>
> "Eckerd and Zenez Recognized by SCO for Longest Running UNIX Servers Eckerd
> runs uninterrupted for a decade on OpenServer, Zenez runs on UNIX for 20
> years"
>
> When SCO announced this contest a few months ago, I thought they were
> looking for the longest *continuously* running SCO server. With all the
> occasional routine maintenance that one has to do on a typical system,
> I knew I didn't have a chance. I expected the winner to have a system
> with a continuous uptime of 2 or 3 years possibly. (No more due to
> the mandatory Y2K upgrades.)
>
> But from the content of this article, it seems that they chose 2 worthy
> companies who have simply *used* SCO systems on a continual basis,
> with breaks for maintenance and upgrades, for 11 and 20 years,
> respectively.
>
> Since my company started with Xenix 2.3.4GT back in 1991, I feel a little
> bit
> left out. Not that I feel that I should have won, because I know that
> there are many people on the ng with years longer experience
> using SCO in a single company.
>
> So did the right companies win this contest in your opinion?
>
> Bob


I read and misunderstood the contest also. I guess that makes Boyd a winner
for getting the rules right.

Anyone care to post the longest uninterrupted up time. At one customers site
a 5.0.4 SMP system was up for just over 400 days, at almost 100% utilization.
It would erase a Progress database, get a data dump from the primary machine,
reload the database, back it up to tape, and run some large reports. Towards
the end it was taking 21.5 to 23 hours everyday at full load to finish. I ran
a "netstat -m" on the machine and it took a minute to realize that the numbers
were so large that the whole screen was garbled, I though the command had gone
bad.

Under normal load I have some customers servers that have run around 700 days,
but the came down for service.

Anyone else?

Mike
--
Michael Brown

The Kingsway Group

> I read and misunderstood the contest also. I guess that makes Boyd a
winner
> for getting the rules right.
>
> Anyone care to post the longest uninterrupted up time. At one customers
site
> a 5.0.4 SMP system was up for just over 400 days, at almost 100%
utilization.
> It would erase a Progress database, get a data dump from the primary
machine,
> reload the database, back it up to tape, and run some large reports.
Towards
> the end it was taking 21.5 to 23 hours everyday at full load to finish.
I ran
> a "netstat -m" on the machine and it took a minute to realize that the
numbers
> were so large that the whole screen was garbled, I though the command had
gone
> bad.
>
> Under normal load I have some customers servers that have run around 700
days,
> but the came down for service.

We had a server that was using a custom written app with roughly 120 users
up for just shy of 500 before bringing it down for maintenance.

Now, I think the best client box we've got out there is 430 days, but
that's on a Linux box.. *hide*

I do have a number of servers which have lost count of their uptime days
(which from memory was a wrap at 248 or something days).

3:50pm up 1 user, load average: 0.03, 0.01, 0.00
4:03pm up 1 user, load average: 0.00, 0.00, 0.00

Longest current uptime on a SCO box:

2:51pm up 493 days, 15:45, 1 user, load average: 0.00, 0.02, 0.00

much fun. These 3 are from 'backup' (DR) machines, thus no users. They do
serve as RDBMS servers for other purposes though, as well as data
mirroring.

some high ranking Linux boxes:

3:04pm up 423 days, 13:34, 4 users, load average: 0.37, 0.34, 0.29
4:17pm up 467 days, 3:28, 1 user, load average: 0.00, 0.00, 0.00
2:48pm up 493 days, 6:22, 1 user, load average: 0.00, 0.01, 0.00

They just keep going.. 'net gateway's and mail/web servers.

fun..

bkx

Mike Brown wrote:
> Anyone care to post the longest uninterrupted up time.

?jul(2003,8,24)-jul(2000,12,10)

987

Client's Red Hat Linux 6.2 box last booted dec 10th, 2000 for 987 days :-)

>
> I had one server run 7 years till a power outage took it down. I used to
> used adb to reset variables or was it kdb. I used a test machine to make
> sure anything I did would not kill machine. It was very fun. It is
> surpising what you will do to keep a system up. I had all kinds of tricks
> that I used to keep client machines up. Weekly logons to reset some
> system parms. I think the longest was 9 years and then the hardware died.
> I had over 50 machines that were up over 3 years. At one of my clients
> there CAD system had never been powered off in 6 years. Most of my
> systems the average is 3-5 years. These are servers not on internet.

Our best was on a V3.2.4.2 ... at a Casino ... somewhere a little over
4 years and 3 months between reboots. We took it down to upgrade the
drives (hard and tape). The drive spun down and locked solid ( ...
we already had our new drive staged and had backed up the data and
transitory directories beforehand.

I heard a story from a good friend to worked for a huge corp in
Australia. They'd run SCO servers for years on the networks .... two
bright recent graduates convinced them to dump the dino and replace
the UNIX with NT. Half way through the multi-million dollar project,
they noticed their reboot schedule wasn't being followed and they
were having good times with NT eating itself. When they explained to
the management that their mission critical servers had to be down or
slowed for rebooting/defragging on a regular basis ...

They got fired. The next team set about reinstalling the new servers
with UNIX.

Moral? Draw your own conclusions. I doubt that Windows and other viri
will ever replace real operating systems like UNIX/Linux.

Jim Hickinbotham
DCRS Solutions

In article <3f499846@dnews.tpgi.com.au>,
Stuart J. Browne <stuart@promed.com.au> wrote:
>> I read and misunderstood the contest also. I guess that makes Boyd a
>> winner for getting the rules right.

>> Anyone care to post the longest uninterrupted up time. At one
>> customers site a 5.0.4 SMP system was up for just over 400
>> days, at almost 100% utilization. It would erase a Progress
>> database, get a data dump from the primary machine, reload
>> the database, back it up to tape, and run some large reports.
>> Towards the end it was taking 21.5 to 23 hours everyday at
>> full load to finish. I ran a "netstat -m" on the machine and
>> it took a minute to realize that the numbers were so large
>> that the whole screen was garbled, I though the command had
>> gone bad.

>> Under normal load I have some customers servers that have run
>> around 700 days, but the came down for service.

>We had a server that was using a custom written app with roughly
>120 users up for just shy of 500 before bringing it down for
>maintenance.

>Now, I think the best client box we've got out there is 430 days, but
>that's on a Linux box.. *hide*

>I do have a number of servers which have lost count of their
>uptime days (which from memory was a wrap at 248 or something
>days).

I thought that was fixed with the 5.0.5 release. I had the same
problem - but ISTR that I had seen the uptime past the 248 day
[that was the lbolt problem as I recall] but when I looked at it
at about 15 months I had not uptimes listed. It finally went down
a month or so later when the power company was out longer than the
UPS could handle it in an overnight situtation.

>some high ranking Linux boxes:

> 3:04pm up 423 days, 13:34, 4 users, load average: 0.37, 0.34, 0.29
> 4:17pm up 467 days, 3:28, 1 user, load average: 0.00, 0.00, 0.00
> 2:48pm up 493 days, 6:22, 1 user, load average: 0.00, 0.01, 0.00

>They just keep going.. 'net gateway's and mail/web servers.

I finally reinstalled the OS in a mail server and a web server
after about 700 days of uptime - as I needed to finally upgrade the
OS. FreeBSD on those.


--
Bill Vermillion - bv @ wjv . com

<snip>

> >I do have a number of servers which have lost count of their
> >uptime days (which from memory was a wrap at 248 or something
> >days).
>
> I thought that was fixed with the 5.0.5 release. I had the same
> problem - but ISTR that I had seen the uptime past the 248 day
> [that was the lbolt problem as I recall] but when I looked at it
> at about 15 months I had not uptimes listed. It finally went down
> a month or so later when the power company was out longer than the
> UPS could handle it in an overnight situtation.

I thought that was fixed also, but *shrug* two boxes showing it. They are
the only two OSR505+RS505A boxes that show it however. Haven't had the
time/opportunity to figure out what's different about 'em.

Known my luck, the utmp/xutmp are just broken causing those details not to
be there any longer.

>
> >some high ranking Linux boxes:
>
> > 3:04pm up 423 days, 13:34, 4 users, load average: 0.37, 0.34, 0.29
> > 4:17pm up 467 days, 3:28, 1 user, load average: 0.00, 0.00, 0.00
> > 2:48pm up 493 days, 6:22, 1 user, load average: 0.00, 0.01, 0.00
>
> >They just keep going.. 'net gateway's and mail/web servers.
>
> I finally reinstalled the OS in a mail server and a web server
> after about 700 days of uptime - as I needed to finally upgrade the
> OS. FreeBSD on those.

nice.. I have to live with managers who think "A quick reboot will fix
it!" is a good diagnostic approach.

bkx