[Info-vax] Using VMS for a web server

[Dirk Munk]

bill at server3.cs.scranton.edu (Bill Gunshannon) wrote:
> In article <5b4d3$5572eb23$5ed4324a$44010 at news.ziggo.nl>,
> 	Dirk Munk <munk at home.nl> writes:
>>
>>
>> I applied for a VMS job several months ago. During the job interview I
>> was asked about my knowledge on all kind of other things, but not on
>> VMS. I wasn't an expert on those areas, so they said that would be a
>> problem. When I replied that they were asking for a VMS expert, they
>> never mentioned those other areas of expertise. Well, that was true, but
>> the VMS system had the habit of doing what it was suppose to do, and it
>> had been doing that for years. It never needed any human assistance or
>> intervention. In fact I would have almost nothing to do if my only work
>> would be maintaining the VMS system.
>
> Not surprising.  A few years ago (the first time I retired from the
> University :-) I applied for a job as a COBOL programmer for the Navy
> down in GA.  Same thing.  After the interview when they asked if I had
> any questions I said the same thing.  "This is supposed to be a COBOL
> Programmer job but you asked no questions about COBOL."  They just kinda
> shrugged it off with a comment of, "I guess you're right."  About four
> months into the jb they started scheduling me to go to courses opn PL/SQL
> and Crystal Reports and other non-COBOL stuff.  And when I asked about it
> the answer was, "Oh, we're getting rid of the COBOL.  We're going to re-
> write it all in PL/SQL and Crystal Reports."  Two months later I left.
> So, did you take the job?  :-)

No, I was over-experienced, I had to much knowledge they told me.

>
>>
>> To be fair, other systems can be made to run very reliable as well, but
>> quite often systems are set up to run, not to run reliable. The drive to
>> build reliable systems often just isn't there.
>
> The only place I hear that is here.  My VMS system (more than a decade ago)
> was more reliable than the one's run by the University data center.  My
> Unix and Windows servers today are v ery reliable.  Moreso than the affore-
> mentioned University data center VMS systems.
>

Let me give you an example. I once was confronted with a cluster of 
three Alpha 8400 systems. Big boys for the time, they had 4GB of memory 
each, and that was a lot in those days. The whole configuration must 
have cost millions. There was group of 10 people running the 
application, solving problems, talking with the customers etc.

They had performance problems with this cluster, I did some basic checks 
and found that the disk cache was still at the default 3MB. They had 
important batch runs that were limited to 2.5MB memory usage. When I 
asked them why on earth they had these ridiculous settings, they didn't 
know.

We moved that application to a much smaller cluster of two ES40's, fixed 
all kind of small problems and stupid settings, and then these 10 people 
could be relocated to other work.

Or that wonderful story of a Solaris server. I walked into a room, and 
saw they were monitoring a system. I asked if there was a problem. Yes 
there was a problem, the performance. They didn't know what to do to 
improve it. They had meetings with the supplier, and the supplier didn't 
know either what to do. So I watched the screen, and saw that only 4GB 
of the 8GB of memory was in use. So I asked about the size of the 
database cache. 350MB they told me. I cursed, and told them to increase 
it to 3GB. Problem solved.

I can go on and on with these stories. People setting up raid 1 sets on 
two partitions of the same disk drive is also a nice one.

Quite often staff doesn't have a clue about how to properly set up a 
system or a database.