Miscellaneous (LONG)

[Steve DeJarnett]

In article <211 at trux.UUCP> car at trux.UUCP (Chris Rende) writes:
>Nixdorf tells me that TOS.3.3 will be out soon (which is OSX 4.4).
>What are the major differences/bug fixes/enhancements/etc... between
>OSX 4.0 and OSX 4.4 (or TOS.3.2 and TOS.3.3).

	4.4 (I think) completes the System V side switch to System V Release 3
(dot something, I think, but I don't recall offhand).

>Are there any plans to move up to System V Release 3?

	Done, if memory serves me correctly.

>What exactly is sendmail and what can it give me?

	sendmail is the MTA (Mail Transfer Agent) from Berkeley.  It's a 
fairly complex piece of software that allows mail routing based on a ruleset.
If you're on the Internet, it's almost a must (although smail 3.1 is supposed
to have an SMTP listener in it).

>Currently, I use SMAIL and PATHALIAS and my System V mailer.

	You can continue to use either or both smail and pathalias (we use
smail in addition to sendmail).  A combination of sendmail and smail or
pathalias gives a fairly full-featured mail delivery system.

>The machine comes with documentation which is System V based.
>Is there any BSD documentation? (Aside from the man sheets).
>Nixdorf tells me they have none. How about Pyramid?

	Pyramid ships a full set of BSD docs.  I don't know if Nixdorf does,
but you should check with them.  If not, you can order a full set of BSD docs
from the Usenix Association (provided you're a member) for something like $60.

>When I run "mailx -s", mailx passes the -s to /bin/mail which doesn't
>like the -s. What is wrong?
>"mailx -s" worked OK on my previous System V box...

	Use /usr/ucb/mail (or /usr/ucb/Mail).  They work with the -s flag.
I don't know that the System V versions will.

>How does the ucb UUCP package compare to the att UUCP package?

	Depends on which you like.  Some people argue that HoneyDanBer (BNU)
UUCP is better than BSD UUCP.  Others say the opposite.  We run BSD, but not
because of any religious beliefs.  Mostly because it's just too much work to
figure out what has to change (probably not much), since we do very little
UUCP these days.

>How does the ucb INIT package compare to the att INIT package?

	Again, user preference, mostly.  We run the AT&T init/getty/login
package simply to get password aging.  If you don't want/need password aging,
you can probably stick with the BSD stuff, but basically, you probably should
go with what you're familiar with.

>I can't get adb or sdb to work at all. I get some type of error message
>about a table being the wrong or old format. dbx does seem to work.

	What universe??  dbx and adb both worked fine for us.  They both
definitely work in the BSD universe of 4.4.

>In order to get shl to work I had to create a /dev/sxt directory and
>"ln /dev/sxt000 /dev/sxt/000" and so on for 001, 002, 003, etc...

	It's there in 4.4 already.

>Here's an odd error message I recieved once while running ps:
>"ps: error locating command name for pid 13216 from Thu Jul 13 14:07@ Q"

	You probably ran ps as a process was exiting and it's U area was
getting cleaned out.  If it happens all the time, you might worry.  If it 
happens once in a blue moon, I wouldn't.  I see this happen on some other
systems much more often (Celerity comes to mind as one).

>What is the "Commercial instuction set" board?

	I'd venture a guess (WARNING:  Only a guess) that this is the stuff
necessary to support the instructions for COBOL that are in the Pyramid
instruction set.

>- Does this machine use virtual memory? Seems not to since I was limited
>  by physical RAM size.

	Yes.  At least Pyramid's do.  I doubt that Nixdorf would rip all of
the VM stuff out.

>- Does this machine do paging of individual RAM pages? Or, is all of the 
>  processes memory swapped in and out?

	Demand paging and swapping.

>"sar" reported an average of "-20647" for kchar/s.

	Many of the System V accounting programs were fixed in 4.4.

>Christopher A. Rende           Central Cartage (Nixdorf/Pyramid/SysV/BSD4.3)

	I hope some of that helps.  I suspect Pyramid people might have more
light to shed on some of the issues you raised.

-------------------------------------------------------------------------------
| Steve DeJarnett            | Smart Mailers -> steve at polyslo.CalPoly.EDU     |
| Computer Systems Lab       | Dumb Mailers  -> ..!ucbvax!voder!polyslo!steve |
| Cal Poly State Univ.       |------------------------------------------------|
| San Luis Obispo, CA  93407 | BITNET = Because Idiots Type NETwork           |
-------------------------------------------------------------------------------