Xref: utzoo comp.sys.hp:9289 comp.unix.aix:5610
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!cs.utexas.edu!uunet!math.fu-berlin.de!unido!uni-koeln!IKP.Uni-Koeln.DE!se
From: se@IKP.Uni-Koeln.DE (Stefan Esser)
Newsgroups: comp.sys.hp,comp.unix.aix
Subject: Re: HP-720 vs IBM-320 vs Sparc2
Message-ID: <1991Jun8.170028.156828@rrz.uni-koeln.de>
Date: 8 Jun 91 17:00:28 GMT
References: <1991Jun6.151807.670@idaho.uucp> <1991Jun7.172642.20985@alchemy.chem.utoronto.ca>
Sender: news@rrz.uni-koeln.de (Usenet News System)
Organization: Institute of Nuclear Physics, University of Cologne, Germany
Lines: 36



|> ourselves have had a lot of problems. The history of HP-UX is much longer,
|> but it is SYS5-based as you pointed out, which is a BIG pain.

Sad to say, at least the HP-PA version of HP-UX is derived from BSD!
(Have a look at the system header files, you'll find lots of BSD 4.3 
'compatible' structures ...). HP sells it as Sys V.2 compatible, but they 
had to remove lots of BSD features to achieve this compatibility!

While the Sys.V ps commands '-u uid' oprion is very nice, most other 
changes to make HP-UX Sys.V compatible several times made me 
quite angry. The print spool system is the worst I've ever seen on Unix.
The spool partitions aren't seperated from root and user file system 
and the spooler has the habit of leaving spool files lying around, if there 
happens not to be enough disc space to hold the whole file!
The user isn't notified, the file isn't removed from the spool dir (thus 
filling the disc further...) and the job will never be printed. 
You probably can imagine what happens every time a new user tries 
to print several big postscript files at once :-(.

We had trouble with just about everything different from a vanilla BSD 
system. (.. archiver with 14 char limit, missing renice(), missing adjtime(), 
missing ulimit() capabilities, missing kernel config. capabilities, missing 
variable sized partitions support and so on :-(  )
(BTW: the NFS problems reported to HP support more than a year ago 
aren'tanswered yet !)

I got to know HP-UX too well and I'm glad we recently were able to get 
a nice, (small, cheap) server machine, that not only performs much better 
than our HP 9000/835, but also doesn't have so braindamaged 'features'.

Stefan
-- 
 Stefan Esser,  Institute of Nuclear Physics,  University of Cologne,  Germany
 se@IKP.Uni-Koeln.DE                                           [134.95.192.50]