[9fans] CPU Server

[9fans] CPU Server

[Franklin]

    Do I need a File Server to install a CPU Server?

Franklin.

Re: [9fans] CPU Server

[Joshua Wood]

I think you are under-reading Erik's comment above about the  
flexibility of these facilities ((/cfg/$sysname/ /bin/)^cpurc[.local]  
and friends) being mere rc. I also think the suggested changes are  
rather narrowly focused on the procedure described in the wiki for  
constructing an `all-in one' cpu/auth/terminal. I'm not sure you've  
given the same consideration to the many ways of distributing Plan 9  
that has been given to the default scripts in the distribution. (This  
is not to say those default scripts can't be improved.)

The suggested changes are really documentation (that's why they are  
all commented) for building a very particular sort of `combo' Plan 9  
system; IMO you should maybe change the wiki but not the scripts.

--
Josh

Re: [9fans] CPU Server

["Rodolfo \"kix\" Garci­a"]

Hola,

IMHO we can make this changes:

1. Create a new file /cfg/example/cpustart with this contents:

# The auth services
# aux/listen -q -t /rc/bin/service.auth -d /rc/bin/service il
# aux/listen -q -t /rc/bin/service.auth -d /rc/bin/service tcp

# To start rio
# aux/mouse $mouseport
# aux/vga -l $vgasize
# exec rio

2. Add this lines to end of the /rc/bin/cpurc.local file

# bind the devices to use
# for (i in m i S t)
#	bind -a '#'^$i /dev >/dev/null >[2=1]


At the other hand, why in the /cfg/example/cpurc file the "dns" command 
is "ndb/dns -s" and not "ndb/dns -r". Probably we need to leave both lines:

# Uncomment if you are a DNS server
# ndb/dns -s
# Uncomment if you are DNS resolver
# ndb/dns -r

After the comments, I will create a patch and I will make the changes in 
the Wiki.

Thanks.

Saludos.

erik quanstrom wrote:
>>> This lines are not in the /cfg/$sysname/cpurc file
>>>
>>>       
>> iirc these lines go into /rc/bin/cpurc, not /cfg/$sysname/cpurc
>>
>>     
>
> actually, neither exactly reflects the state of the current dist's cpurc.
> there are provisions for "site-specific startup", "cpu-specific startup"
> (/cfg/$sysname/cpurc) and "cpu-specific late startup"
> (/cfg/sysname/cpustart).
>
> but since it's just a shell script, you can change it any way you'd like.
> i find it a bit easier to manage changes with big switch statement.
> a vast majority of our cpu servers don't have a case in the switch
> anyway.
>
> but then again, i'm a well-known luddite. ;-)
>
> - erik
>   

Re: [9fans] CPU Server

[erik quanstrom]

> > 
> > This lines are not in the /cfg/$sysname/cpurc file
> > 
> 
> iirc these lines go into /rc/bin/cpurc, not /cfg/$sysname/cpurc
> 

actually, neither exactly reflects the state of the current dist's cpurc.
there are provisions for "site-specific startup", "cpu-specific startup"
(/cfg/$sysname/cpurc) and "cpu-specific late startup"
(/cfg/sysname/cpustart).

but since it's just a shell script, you can change it any way you'd like.
i find it a bit easier to manage changes with big switch statement.
a vast majority of our cpu servers don't have a case in the switch
anyway.

but then again, i'm a well-known luddite. ;-)

- erik

Re: [9fans] CPU Server

[John Soros]

Hi!
There definitely is some truth about the wiki page not being completely correct, though it is not very hard to figure out either.

> Hi,
> 
> the page Configuring_a_Standalone_CPU_Server is not correct and the 
> people can have problems to configure a CPU Server. Examples:
> 
> 1 - - -
> 
> "cd /cfg; mkdir $sysname; dircp example $sysname
> (...)
> At the end of cpurc.local ..."
> 
> The cpurc.local is only in /bin, not in /cfg/example/
> 
> 2 - - -
> 
> 'cat /dev/drivers' will list the available devices. In this case we have 
> m (mouse), i (draw), S (sd - disk), and t (uart - serial); if you get 
> errors about /dev/realmode, include P in this list:
> 
> for (i in m i S t)
>     bind -a '#'^$i /dev >/dev/null >[2=1]
> 
> Is OK, but ... why is not it in the file like in /bin/termrc?
> 
> 3 - - -
> 
> Uncomment the two lines indicated in cpurc to enable the authentication 
> functions:
> 
> # auth/keyfs -wp -m /mnt/keys /adm/keys >/dev/null >[2=1]
> # auth/cron >>/sys/log/cron >[2=1] &
> 
> This lines are not in the /cfg/$sysname/cpurc file
> 

iirc these lines go into /rc/bin/cpurc, not /cfg/$sysname/cpurc

> 4 - - -
> 
> Add these lines to /cfg/$sysname/cpustart
> 
> The cpustart file do not exists
> 
>  - - -
> 
> Thanks.

It would be great if you could change the wiki page!
Cheers!

Re: [9fans] CPU Server

[Iruata Souza]

On Dec 22, 2007 9:12 PM, "Rodolfo \"kix\" Garci­a" <kix@kix.es> wrote:
> Hi,
>
> the page Configuring_a_Standalone_CPU_Server is not correct and the
> people can have problems to configure a CPU Server. Examples:
>
> 1 - - -
>
> "cd /cfg; mkdir $sysname; dircp example $sysname
> (...)
> At the end of cpurc.local ..."
>
> The cpurc.local is only in /bin, not in /cfg/example/
>
> 2 - - -
>
> 'cat /dev/drivers' will list the available devices. In this case we have
> m (mouse), i (draw), S (sd - disk), and t (uart - serial); if you get
> errors about /dev/realmode, include P in this list:
>
> for (i in m i S t)
>     bind -a '#'^$i /dev >/dev/null >[2=1]
>
> Is OK, but ... why is not it in the file like in /bin/termrc?
>
> 3 - - -
>
> Uncomment the two lines indicated in cpurc to enable the authentication
> functions:
>
> # auth/keyfs -wp -m /mnt/keys /adm/keys >/dev/null >[2=1]
> # auth/cron >>/sys/log/cron >[2=1] &
>
> This lines are not in the /cfg/$sysname/cpurc file
>
> 4 - - -
>
> Add these lines to /cfg/$sysname/cpustart
>
> The cpustart file do not exists
>

you seem correct to me. if nobody complains, you could change the wiki.

iru

[9fans] CPU Server

["Rodolfo \"kix\" Garci­a"]

Hi,

the page Configuring_a_Standalone_CPU_Server is not correct and the 
people can have problems to configure a CPU Server. Examples:

1 - - -

"cd /cfg; mkdir $sysname; dircp example $sysname
(...)
At the end of cpurc.local ..."

The cpurc.local is only in /bin, not in /cfg/example/

2 - - -

'cat /dev/drivers' will list the available devices. In this case we have 
m (mouse), i (draw), S (sd - disk), and t (uart - serial); if you get 
errors about /dev/realmode, include P in this list:

for (i in m i S t)
    bind -a '#'^$i /dev >/dev/null >[2=1]

Is OK, but ... why is not it in the file like in /bin/termrc?

3 - - -

Uncomment the two lines indicated in cpurc to enable the authentication 
functions:

# auth/keyfs -wp -m /mnt/keys /adm/keys >/dev/null >[2=1]
# auth/cron >>/sys/log/cron >[2=1] &

This lines are not in the /cfg/$sysname/cpurc file

4 - - -

Add these lines to /cfg/$sysname/cpustart

The cpustart file do not exists

 - - -

Thanks.

Re: [9fans] CPU Server

[erik quanstrom]

> Hello,
> 
> I have a question regarding cpurc. I managed to set up a combined cpu/auth
> server by loosely following the wiki instructions.
> 
> If I understand correctly, the lines
> 
> if(! test -e /rc/bin/service.auth/il566){
> 	mv /rc/bin/service.auth/authsrv.il566 /rc/bin/service.auth/il566
> 	mv /rc/bin/service.auth/authsrv.tcp567 /rc/bin/service.auth/tcp567
> 	mv /rc/bin/service/il566 /rc/bin/service/_il566
> 	mv /rc/bin/service/tcp567 /rc/bin/service/_tcp567
> }
> 
> make the auth server listen on the right ports. Is there a likely scenario
> where these will be executed again after the first time the server boots, or
> can I remove them?
> 
> Please let me know if I'm missing something obvious. I'm quite new to Plan 9.
> 
> Cheers
> Rob

you're supposed to do that once.

this is what i have at home

	cpu% lc /rc/bin/service.auth
	!tcp110	il566	tcp22	tcp566	tcp567	tcp995

you may not want ssh.  to keep ssh from starting

	cpu% mv tcp22 !tcp22

of course, i don't have il567 or tcp567 in /rc/bin/service.
just rm them of they're there.

to turn all these  services on, you need to listen for them
from cpurc.  these are the last two lines of mine

	aux/listen -q -t /rc/bin/service.auth -d /rc/bin/service il
	aux/listen -q -t /rc/bin/service.auth -d /rc/bin/service tcp

though if you don't use ken's fileserver, you don't need the
first line as you don't need il.

- erik

Re: [9fans] CPU Server

[rob]

Hello,

I have a question regarding cpurc. I managed to set up a combined cpu/auth
server by loosely following the wiki instructions.

If I understand correctly, the lines

if(! test -e /rc/bin/service.auth/il566){
	mv /rc/bin/service.auth/authsrv.il566 /rc/bin/service.auth/il566
	mv /rc/bin/service.auth/authsrv.tcp567 /rc/bin/service.auth/tcp567
	mv /rc/bin/service/il566 /rc/bin/service/_il566
	mv /rc/bin/service/tcp567 /rc/bin/service/_tcp567
}

make the auth server listen on the right ports. Is there a likely scenario
where these will be executed again after the first time the server boots, or
can I remove them?

Please let me know if I'm missing something obvious. I'm quite new to Plan 9.

Cheers
Rob

Re: [9fans] CPU Server

[geoff]

I've updated that section of the Wiki to reflect current reality.

[9fans] CPU Server

[Rodolfo (kix)]

Hi,

I am trying to install a new CPU Server, but the wiki document is wrong.

In http://cm.bell-labs.com/wiki/plan9/Configuring_a_standalone_CPU_server/index.html

we can read:

At the top of the file, just after the facedom= line

But, this line is in the cpurc.local file.

I found this patch, probably the origin of the changes:

http://plan9.cs.bell-labs.com/sources/extra/changes/2007/0413

But I am not sure to do any changes in the cpurc file. Any help?

Thanks.

-- 
Rodolfo García "kix"
EA4ERH - IN80ER

Re: [9fans] cpu server

[Jack Johnson]

On Feb 9, 2004, at 9:00 PM, Geoff Collyer wrote:

> Canada is large and varied.  Temperatures in the summer around 40°C
> are possible on the prairies and in Ontario, at least.  Winters tend
> to run the gamut from bitterly cold to damned cold, though there are
> exceptions, such as Vancouver and Victoria, where it rarely reaches
> freezing.  Roughly 90% of the population live within 160km (100 miles)
> of the US border; as you get farther north, it gets colder, and Canada
> extends up to the north pole.  Currently (~20:30 Vancouver time), it's
> a balmy 5°C in Vancouver (and Calgary, amazing!), 2°C in Montreal,
> -30°C in Alert, Nunavut, -39°C in Tuktoyaktuk, NWT, but that doesn' t
> include the wind chill.

I went to college in interior Alaska which is very similar to interior 
Canadian weather.  I've seen -68F without wind chill in the winter 
(though that's the extreme) and 96F in the summer, all in the same 
city.

But again, so large that when people say, "Oh, yeah, I know, I lived in 
Southeast," I just chuckle, because you might as well live in Victoria 
or Vancouver (or Bellingham!), and Fairbanks weather is nothing 
compared to anything on the Arctic coast mid-winter.

Fort St. John, now there I can sympathize....

-Jack

Re: [9fans] cpu server

[Geoff Collyer]

Canada is large and varied.  Temperatures in the summer around 40°C
are possible on the prairies and in Ontario, at least.  Winters tend
to run the gamut from bitterly cold to damned cold, though there are
exceptions, such as Vancouver and Victoria, where it rarely reaches
freezing.  Roughly 90% of the population live within 160km (100 miles)
of the US border; as you get farther north, it gets colder, and Canada
extends up to the north pole.  Currently (~20:30 Vancouver time), it's
a balmy 5°C in Vancouver (and Calgary, amazing!), 2°C in Montreal,
-30°C in Alert, Nunavut, -39°C in Tuktoyaktuk, NWT, but that doesn' t
include the wind chill.

If we're smart, we'll go ahead with the annexation of the Turks and
Caicos Islands this time.

Even in Japan you must have quite a range of temperature; it snows on
Hokkaido, doesn't it?

Re: [9fans] cpu server

[Lyndon Nerenberg]

--On 2004-2-9 6:11 PM -0700 andrey mirtchovski
<mirtchov@cpsc.ucalgary.ca> wrote:

> Just couple of weeks ago the temperature measured the same number in
> both Celsius and Fahrenheit...

The same week that it was -40 C/F in Edmonton, Alberta, Canada, it was
+8C (say 45F) here in Vancouver, British Columbia, Canada.

Can you guess why I'm in the process of moving from Edmonton to
Vancouver? ;-)

--lyndon

Re: [9fans] cpu server

[boyd, rounin]

> I've thought Canada is more cold than that!
> I was wrong.

depends where you are.  it can go seriously negative in Montréal.

Re: [9fans] cpu server

[David Presotto]

[-- Attachment #1: Type: text/plain, Size: 21 bytes --]

That was a hot day...

[-- Attachment #2: Type: message/rfc822, Size: 1985 bytes --]

From: Kenji Okamoto <okamoto@granite.cias.osakafu-u.ac.jp>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] cpu server
Date: Tue, 10 Feb 2004 10:38:15 +0900
Message-ID: <c9a1d231132442d5599098ee715ea99e@granite.cias.osakafu-u.ac.jp>

>right now I must go ice skating -- this
> is canada after all (it's -1 degrees centigrade and people are wearing
> shorts).

I've thought Canada is more cold than that!
I was wrong.

Kenji

Re: [9fans] cpu server

[Kenji Okamoto]

>right now I must go ice skating -- this
> is canada after all (it's -1 degrees centigrade and people are wearing
> shorts).

I've thought Canada is more cold than that!
I was wrong.

Kenji

Re: [9fans] cpu server

[andrey mirtchovski]

>> I've thought Canada is more cold than that!
>> I was wrong.

Just couple of weeks ago the temperature measured the same number in
both Celsius and Fahrenheit...

"It was SO cold, some people in the US were seriously considering
whether to count Canadian weather masses as WMDs."

thank you... thank you... i'm here every tuesday!

Re: [9fans] cpu server

[9nut]

> right now I must go ice skating -- this
> is canada after all (it's -1 degrees centigrade and people are wearing
> shorts).

Plan9 on Ice? ☺

Re: [9fans] cpu server

[andrey mirtchovski]

>> but I can sacrifice my university bandtwidth for its weekly update.
>
> to make it useful, it has to have a much finer granularity than that,
> preferably instantaneous or on the order of minutes, then we can
> play edit wars a la wikipedia...
>
> http://meta.wikipedia.org/wiki/Edit_wars

sure, run it daily, if BL doesn't mind us soaking up the connection...

Re: [9fans] cpu server

[andrey mirtchovski]

>> put the wiki in a replica and you'll get updates unformation
>> automagically :)
>
> well... if wikifs provided date stamps.
>
>> I'm looking at the wikifs right now and it looks like there's enough
>> timestamps there to make such a script easy to write.
>
> to me, it looks as if wikifs doesn't maintain modification times
> or Qid version numbers.
>
> maybe that's the easiest modification to make;
> then a simple "ls -lt" would go a long way...

not at the top-level directory but on the ones underneath it does:

this ugly piece tells me the directories changed in the past week:

	#!/bin/rc

	rfork en
	ramfs

	test -e /srv/wiki || srv 'net!plan9.bell-labs.com!wiki' wiki || exit

	mount -c /srv/wiki /mnt/wiki
	cd /mnt/wiki

	for (i in *) {
		for(j in $i/*) {
				if(test '(' -d $j ')' -a '(' ! $j -older 7d ')')
					echo $i # or echo $j to see the actual timestamp
			}
		}
	}

if you run it you'll see at least two (so far) -- [dj]'s_jedimaters and grumpybear

both have nothing to do with plan9 :)

Re: [9fans] cpu server

[rog]

> but I can sacrifice my university bandtwidth for its weekly update.

to make it useful, it has to have a much finer granularity than that,
preferably instantaneous or on the order of minutes, then we can
play edit wars a la wikipedia...

http://meta.wikipedia.org/wiki/Edit_wars

Re: [9fans] cpu server

[andrey mirtchovski]

>> right now I must go ice skating -- this
>> is canada after all (it's -1 degrees centigrade and people are wearing
>> shorts).
> 
> Plan9 on Ice? ☺

glenda, rather :)

Re: [9fans] cpu server

[rog]

> put the wiki in a replica and you'll get updates unformation
> automagically :)

well... if wikifs provided date stamps.

> I'm looking at the wikifs right now and it looks like there's enough
> timestamps there to make such a script easy to write.

to me, it looks as if wikifs doesn't maintain modification times
or Qid version numbers.

maybe that's the easiest modification to make;
then a simple "ls -lt" would go a long way...

Re: [9fans] cpu server

[andrey mirtchovski]

> i've sometimes thought that the lack of an easy "what's new" page
> makes a difference.  ok, i've been too lazy to write a script to do
> it, but in my defense, such a script would probably have to traverse
> the entire wiki each time, which wouldn't be great if lots of people
> were doing it, and certainly wouldn't be quick.
>

put the wiki in a replica and you'll get updates unformation
automagically :)

I'm looking at the wikifs right now and it looks like there's enough
timestamps there to make such a script easy to write.  maybe not a
good idea to run it from the UK, but I can sacrifice my university
bandtwidth for its weekly update.

andrey

ps: i've started on it, will send it to you later tonight for
review if you're interested.  right now I must go ice skating -- this
is canada after all (it's -1 degrees centigrade and people are wearing
shorts).

Re: [9fans] cpu server

[rog]

> I think the fundamental failure of the Plan 9 wiki is because it has
> (always) been superseded by this list.  As it currently is, it needs a
> huge amount of cleaning up -- it's filled with random pages, some even
> may have questionable content.

i've sometimes thought that the lack of an easy "what's new" page
makes a difference.  ok, i've been too lazy to write a script to do
it, but in my defense, such a script would probably have to traverse
the entire wiki each time, which wouldn't be great if lots of people
were doing it, and certainly wouldn't be quick.

with a page that tells you quickly and easily what's changed, people
can make a habit of having a brief sanity check of any new stuff that's been
added, and maybe add a tidbit or two if so inclined.

without it, it's easy just to look at the front page, see nothing different,
and leave it at that.

that's what i do, i'm afraid.

Re: [9fans] cpu server

[Bruce Ellis]

I just put a sticker on my plan 9 machine saying "Jim says you are brain dead.".
Then I checked that it still had blue wires hanging out it's rear and that it was
happily talking to the rest of the pad.  It seems to have completed it's over-
night tasks and everyone is happy.  I don't think I'll believe Jim.

I'm certainly not going to put it in a cloud, it looks like rain today.

I'm not sure it took offence but I won't show it this ...

> http://open-forge.com/hangar18/Plan9Intro.txt
>
> This document explains how to do a single I/O,cpu,auth server using kfs.

It's set up better than that example and it will only gloat.
Anyway it has a little hangar now so I don't have to read Jim's mail.

brucee

Re: [9fans] cpu server

[Bruce Ellis]

I guess you didn't read my maill.  I have a single plan 9 box and
I use it to great advantage.  You might like to repeat writing the
same erroneous statement again and again.

brucee

> That is correct Bruce. A single P9 box is brain dead, it is unable to
> demonstrate -any- of its advantages over a traditional OS.

Re: [9fans] cpu server

[Jim Choate]

On Sat, 7 Feb 2004, Charles Forsyth wrote:

> >>That is correct Bruce. A single P9 box is brain dead, it is unable to
> >>demonstrate -any- of its advantages over a traditional OS.
>
> actually, i don't think that's quite right, if you look at the construction
> of many of the services, for instance.

Yada yada yada, backpedaling.

P9 -is- a *distributed operating system*. That was it's purpose. Using it
for anything else, while possible, isn't using its strengths. It's simply
trying to shoehorn it in to the same old same old. That doesn't do P9 the
justice it deserves. It does of course explain why the OS has not grown
with the potential it possess. If you don't see, you don't see.

You're of course welcome to your own opinion.

 -- --

Open Forge, LLC  24/365 Onsite Support for PCs, Networks, & Game Consoles
512-695-4126 (Austin, Tx.)  help@open-forge.com  irc.open-forge.com

Hangar 18  Open Source Distributed Computing Using Plan 9 & Linux
512-451-7087  http://open-forge.org/hangar18  irc.open-forge.org

James Choate  512-451-7087  ravage@ssz.com  jchoate@open-forge.com

Re: [9fans] cpu server

[Taj Khattra]

On Sat, Feb 07, 2004 at 05:26:33PM -0500, a@9srv.net wrote:
> wow, charles. excelent write-up.

yes, very nice

> lines of shell code. the amount of times since then that i've
> wished for aux/listen on whatever unix i was working on at the
> time is beyond number. it's not just incremental; it's totally

tcpserver on unix gives you something sort of similar to aux/listen

the tcp7 equivalent is

	tcpserver -RHl0 0 7 /bin/cat

the tcp9 equivalent is

	tcpserver -RHl0 0 9 sh -c 'exec /bin/cat >/dev/null'

-taj

Re: [9fans] cpu server

[andrey mirtchovski]

> wow, charles. excelent write-up. just a few additions i'm
> compelled to make, based on recent experience and first impressions.

alright, I put both your comments on the wiki:

http://plan9.bell-labs.com/wiki/plan9/what_do_people_like_about_plan_9/index.html

I think the fundamental failure of the Plan 9 wiki is because it has
(always) been superseded by this list.  As it currently is, it needs a
huge amount of cleaning up -- it's filled with random pages, some even
may have questionable content.

Re: [9fans] cpu server

[a]

wow, charles. excelent write-up. just a few additions i'm
compelled to make, based on recent experience and first impressions.

when i was first exposed to plan 9, i was a unix admin. the very
first thing that struck me about it was /lib/ndb. plan 9 was
build with networks in mind, and as such the handling of them is so
far superior to, well, everything else as to be operating at a
different level entirely. no more changing host names in 3-5
different places, no more worrying about if my dns data was synched
with my arp data and so on. designing the system around the idea of
being network-centric meant dealing with and organizing a network
was simple, even when not making use of the advanced networking
capabilities the system presents. the small stuff falls out of doing
doing the big stuff right.

a few days later i had my second big "whoa" moment with plan 9 while
while looking at aux/listen. tcp7 and tcp9 just blew my mind. i
could write network listeners! in like three lines!!! wereas
something like xinetd is (arguably; or not) an incremental step up
from inetd on modern unix systems, aux/listen is something else
entirely. i've written cross-protocol port forwarding in two
lines of shell code. the amount of times since then that i've
wished for aux/listen on whatever unix i was working on at the
time is beyond number. it's not just incremental; it's totally
different. and it works stand-alone, too.

again: the small stuff falls out of doing the big stuff right.

and plan 9 does the big stuff right.
ア

Re: [9fans] cpu server

[Charles Forsyth]

>>That is correct Bruce. A single P9 box is brain dead, it is unable to
>>demonstrate -any- of its advantages over a traditional OS.

actually, i don't think that's quite right, if you look at the construction
of many of the services, for instance.  initial design is often
given a focus by thinking about possible use of the name space and
then the design or structure of that name space; when it makes sense,
the consequent split into name-space provider and name-space user(s)
can simplify application structure, regardless whether distribution is
involved or not.  thus, there is still advantage in its use and significant
difference with more `traditional' systems.

now, a message-passing system (say) will obviously encourage
applications to be based on message-passing, but plan 9's name space
sits at a higher level of abstraction than message-passing (or soft
SOAP), and therefore comes in at a higher level of application design
(i'd say).   use of name spaces is encouraged and
supported by mechanisms deeply embedded into the system itself.

acme provides a good example of using those ideas on a single machine.
it's a file server but that's not primarily for reasons of
distribution, and indeed i haven't seen much use of its services being
subject to import or export (though they could be).  its activity is
usually confined to a single box, but it's still an advantageous
structure.  that's why i often call it an `integrating' not an
`integrated' environment.  the latter tend to be self-contained (i
know, i know.  `plug-ins', where you have to come to grips with deep
knowledge of the IDE's internals).  acme is much more open-ended, and
i shouldn't think the typical acme client has any idea or need to know
what acme looks like inside.

keyfs and factotum similarly simplify the interface for their clients,
by being name space servers, again in a different way from traditional OSs.
(implementation of the service itself is also given focus by its being a file server.)

of course, you're right that there is still greater power once the name
space is available across a network.
one of the interesting things about the Vita grid work to me
was how representing things in the name space
led to a simple yet flexible interface from the client's point of view.
it's described in terms of reading and writing a few files,
and it's natural to describe the protocol for the use of those files.
i wasn't involved in the design or implementation, but
i found it easy to see what a new client had to do,
much more so than with the more traditional API/ABI descriptions.

Re: [9fans] cpu server

[Jim Choate]

On Sun, 8 Feb 2004, Bruce Ellis wrote:

> you said brain-dead.
>
> > I didn't say you couldn't do it, I'm simply saying it brings you -nothing-
> > over -any other OS-.

That is correct Bruce. A single P9 box is brain dead, it is unable to
demonstrate -any- of its advantages over a traditional OS.

-The- point of using a OS is that it brings you something that others
can't. Otherwise it's not a matter of technical supperiority but 'taste'.
And for that there is no accounting.

 -- --

Open Forge, LLC  24/365 Onsite Support for PCs, Networks, & Game Consoles
512-695-4126 (Austin, Tx.)  help@open-forge.com  irc.open-forge.com

Hangar 18  Open Source Distributed Computing Using Plan 9 & Linux
512-451-7087  http://open-forge.org/hangar18  irc.open-forge.org

James Choate  512-451-7087  ravage@ssz.com  jchoate@open-forge.com

Re: [9fans] cpu server

[Bruce Ellis]

you said brain-dead.

> I didn't say you couldn't do it, I'm simply saying it brings you -nothing-
> over -any other OS-.

Re: [9fans] cpu server

[Jim Choate]

On Sat, 7 Feb 2004, andrey mirtchovski wrote:

> Dharani,
>
> nevermind choate -- he's high on rage, low on useful information.
>
> there's no reason why you can't boot your cpu server as yourself and
> not as bootes.  I ran a standalone P9 cpu/auth/kfs installation booted
> as 'andrey' for a year in 2000-2001.

I didn't say you couldn't do it, that's -not- what he's asking.

What he wants to do is log in as somebody -other- than bootes. The
canonical way P9 is -designed- to be used.

Boot box as bootes, use it as another user.

He's asking how, the dox as provied on the P9 site -suck- in this regards.

Just like another response, you're changing the rules in the middle of the
game to suite -your- view.

Typical.

 -- --

Open Forge, LLC  24/365 Onsite Support for PCs, Networks, & Game Consoles
512-695-4126 (Austin, Tx.)  help@open-forge.com  irc.open-forge.com

Hangar 18  Open Source Distributed Computing Using Plan 9 & Linux
512-451-7087  http://open-forge.org/hangar18  irc.open-forge.org

James Choate  512-451-7087  ravage@ssz.com  jchoate@open-forge.com

Re: [9fans] cpu server

[Jim Choate]

On Sat, 7 Feb 2004, Bruce Ellis wrote:

> Maybe I'll be the only one to bother arguing with Jim,
> but this comment is just a waste of time.
>
> > A single P9 box is truly brain-dead [1].
>
> I usually only run one P9 box, as I'm doing now.
> It's far from brain-dead and hasn't crashed yet, ever.
> To use it's resources from afar, normally the couch,
> I can drawterm or ssh to it from almost anything.

I didn't say you couldn't do it, I'm simply saying it brings you -nothing-
over -any other OS-.

As to your commentary about crashing and stuff, I made no such statement.
You are changing the rules of the game right in the middle.

Typical.

 -- --

Open Forge, LLC  24/365 Onsite Support for PCs, Networks, & Game Consoles
512-695-4126 (Austin, Tx.)  help@open-forge.com  irc.open-forge.com

Hangar 18  Open Source Distributed Computing Using Plan 9 & Linux
512-451-7087  http://open-forge.org/hangar18  irc.open-forge.org

James Choate  512-451-7087  ravage@ssz.com  jchoate@open-forge.com

Re: [9fans] cpu server

[boyd, rounin]

> nevermind choate -- he's high on rage, low on useful information.

iirc [la] rage in french == rabies.

http://www.pasteur.fr/recherche/rage

Re: [9fans] cpu server

[boyd, rounin]

> nevermind choate -- he's high on rage, low on useful information.

from https://club-birriga.kicks-ass.org/studio/server_status.asp

[the raw text will ensure is looks dreadful, unlike the real thing]

----


Server Status

----------------------------------------------------------------------------
----

Status:   Sat, 07 Feb 2004 13:15:09 GMT


----------------------------------------------------------------------------
----

Process Status

      Pid Pgrp User State Mem Module
      1 1 bootes release 1K Sh[$Sys]
      9 1 bootes release 3K Styxlib[$Sys]
      10 1 bootes recv 16K MemFS
      13 1 bootes release 1K WebServer[$Sys]
      14 1 bootes release 1K WebServer[$Sys]
      15 15 bootes recv 27K SSLHS
      149 1 bootes ready 3K Asp[$Sys]
      150 1 bootes release 3K HIO[$Sys]



      User:  boyd

Local: 10.0.0.100!443  - Remote: 81.64.132.103


----------------------------------------------------------------------------
----

© 2003, Bruce Ellis

Re: [9fans] cpu server

[a]

// ...bootes is just another username that sounds funny
// when you say it out loud :)

hah! you think "bootes" sounds funny? when setting up my first
personal 3e network, i wanted to convince myself that bootes
wasn't special somewhere (other than /lib/ndb/auth):

:;ps | sed 5q
xyzzy             1    0:00   0:00      88K Await    init
xyzzy             2   95:15   0:00       0K Wakeme   genrandom
xyzzy             3    0:00   0:00       0K Wakeme   alarm
xyzzy             4    0:00   0:00       0K Wakeme   rxmitproc
xyzzy             5    0:00   0:00       0K Wakeme   #l0watchdog

we don't even have a bootes on the system.
ア

Re: [9fans] cpu server

[Bruce Ellis]

cool, party on.

----- Original Message -----
From: "andrey mirtchovski" <mirtchov@cpsc.ucalgary.ca>
To: <9fans@cse.psu.edu>
Sent: Saturday, February 07, 2004 7:18 PM
Subject: Re: [9fans] cpu server


> Dharani,
>
> nevermind choate -- he's high on rage, low on useful information.

Re: [9fans] cpu server

[andrey mirtchovski]

Dharani,

nevermind choate -- he's high on rage, low on useful information.

there's no reason why you can't boot your cpu server as yourself and
not as bootes.  I ran a standalone P9 cpu/auth/kfs installation booted
as 'andrey' for a year in 2000-2001.

just to make sure, I rebooted one of my servers with my authid (I have
no special privileges on the system and am not a member of any groups)
and I'm typing this having drawterm-ed to it:

plan9-2% ps
andrey            1    0:00   0:00      92K Await    init
andrey            2    0:59   0:00       0K Wakeme   genrandom
andrey            3    0:00   0:00       0K Wakeme   alarm
andrey            4    0:00   0:00       0K Wakeme   rxmitproc
andrey            6    0:00   0:00       0K Wakeme   etherread4
andrey            7    0:00   0:00       0K Wakeme   recvarpproc
andrey            8    0:00   0:00       0K Wakeme   etherread6
andrey            9    0:00   0:00     164K Sleep    ipconfig
andrey           11    0:00   0:00       0K Wakeme   #I0tcpack
andrey           12    0:00   0:00     268K Pread    factotum
andrey           13    0:00   0:00       0K Idle     pager
andrey           24    0:00   0:00     164K Sleep    ipconfig
andrey           27    0:00   0:00     172K Pread    cs
andrey           31    0:00   0:00     304K Pread    dns
andrey           40    0:00   0:00      92K Sleep    timesync

so, there.  I can cpu out, I can drawterm and cpu in.  with more users
it just gets a bit more tricky because you can get their passwords in
plain text from keyfs.

remember -- there's no superuser in plan9, bootes is just another
username that sounds funny when you say it out loud :)

andrey

Re: [9fans] cpu server

[Bruce Ellis]

Maybe I'll be the only one to bother arguing with Jim,
but this comment is just a waste of time.

> A single P9 box is truly brain-dead [1].

I usually only run one P9 box, as I'm doing now.
It's far from brain-dead and hasn't crashed yet, ever.
To use it's resources from afar, normally the couch,
I can drawterm or ssh to it from almost anything.

brucee

[9fans] cpu server

[vdharani]

hi,

this may be a silly question.

i have setup a cpu server. i also enabled graphics and started rio. i am
wondering if i can use it as a normal user (say, glenda or myself) while it
acts as a server for other machines. how do i do it? also, if i need to
switch user, how can i do it. i am looking for something like login/logout
or atleast onetime login as someother user (other than bootes).

any help?

thanks
dharani

Re: [9fans] cpu server

[Jim Choate]

Hi Dharani,

On Fri, 6 Feb 2004 vdharani@infernopark.com wrote:

> i have setup a cpu server. i also enabled graphics and started rio. i am
> wondering if i can use it as a normal user (say, glenda or myself) while it
> acts as a server for other machines. how do i do it? also, if i need to
> switch user, how can i do it. i am looking for something like login/logout
> or atleast onetime login as someother user (other than bootes).

This is one of the big differences in thought process that P9 takes over
traditional OSes. Don't confuse who boots the server with users. Under P9
(it being a distributed system) it is -not- meant that users actually sit
at the console to access resources. The console is god knows where and
you're using a I/O server to connect from some completely otherwhere. A
user of P9 resources is -assumed- to be using a I/O server to access other
resources. The setup the vast majority of us have is not 'normal'.

The reality is that P9 assumes as part of its base design you have access
to multiple machines. A single P9 box is truly brain-dead [1].

Right now  you've got three choices as I understand it:

1. From your shell ssh to the appropriate machine (even if it's a single
   one) and log in as a user who has an account.

2. Use drawterm from other machines that are not P9 and log in as that
   user.

3. Use a second machine configured so that it looks for auth from the same
   source as the machines you want to log in on. From there you can use
   attach() and bind() to manage your namespaces.

In any case you must have a auth server running that is shared between the
various resources and configured for those additional users.

[1] And yet the 'traditional' P9 community (ie this list and the OS
    developers) have -never- made a concerted attempt at fostering the
    growth of such resources. This frustration on my part caused me to
    take Hangar 18 on a seperate path. In a very real way they make
    promises and then never deliver.

 -- --

Open Forge, LLC  24/365 Onsite Support for PCs, Networks, & Game Consoles
512-695-4126 (Austin, Tx.)  help@open-forge.com  irc.open-forge.com

Hangar 18  Open Source Distributed Computing Using Plan 9 & Linux
512-451-7087  http://open-forge.org/hangar18  irc.open-forge.org

James Choate  512-451-7087  ravage@ssz.com  jchoate@open-forge.com