On Sat, 21 Oct 2000 18:38:33 -0700
Jordan Hubbard [EMAIL PROTECTED] said:
jkh Sounds good to me. My comments were, just to make it clear again,
jkh just food for thought and not out-and-out objections. If even 47 more
jkh files in /etc is what it takes to get IPv6 fully supported, then so be
There are many discussion aboud having NetBSD style rc.d. However, I
think it takes for a period of time.
Once, I wish to commit my changes to be in time for 4.2-RELEASE.
I think people were talking only about -current here anyway.
A NetBSD style rc.d is certainly not planned for -stable.
-
On Tue, Oct 24, 2000 at 02:56:07PM -0700, Jordan Hubbard wrote:
[redirected to just -current; I'm not sure what this has to do with -net]
I agree. I've been using them for a while on my dog slow Windows CE
machine. There were some minor issues when they were first committed
to NetBSD
In message [EMAIL PROTECTED] Nik Clayton writes:
: On Tue, Oct 24, 2000 at 02:56:07PM -0700, Jordan Hubbard wrote:
: [redirected to just -current; I'm not sure what this has to do with -net]
:
: I agree. I've been using them for a while on my dog slow Windows CE
: machine. There were
On Wed, Oct 25, 2000 at 15:17 -0500, Mike Meyer wrote:
Gerhard Sittig writes:
What's new is:
- include the general config at the start (and yes, in every
single script -- but this should be neglectable in terms of
speed penalty and makes them work separately, too -- which is a
2 ;-)
Bye,
Andrea
-Original Message-
From: David O'Brien [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, October 24, 2000 10:27 PM
To: Warner Losh
Cc: [EMAIL PROTECTED]
Subject: Re: new rc.network6 and rc.firewall6
On Tue, Oct 24, 2000 at 12:31:57PM -0600, Warner Losh wrote:
T
This was my thought also. I put the TCP/IP scripts at 99 to make
sure that any slow network initialization is done.
Since they all start with S - for example S99tcp - moving it
to s99tcp will keep it from starting, and the Knnname in the same
directory is used to stop things when moving
On Wed, Oct 25, 2000 at 06:04 +0700, Alexey Dokuchaev wrote:
On Tue, 24 Oct 2000, David O'Brien wrote:
On Tue, Oct 24, 2000 at 04:23:40PM +0700, Alexey Dokuchaev wrote:
Why can't I simply write kill -1 `cat
/var/run/sendmail.pid`?
What about deamons that don't understand `kill
On Tue, Oct 24, 2000 at 14:56 -0700, Jordan Hubbard wrote:
[ ... NetBSD (or Linux?) like rc scripts ... ]
So, who wants to do a proof-of-concept implementation for
-current which integrates with our existing rc.conf mechanism?
In order to obey POLA, we should at least have the separate
Gerhard Sittig writes:
On Wed, Oct 25, 2000 at 06:04 +0700, Alexey Dokuchaev wrote:
Though I see your point, actually, many UNIX books, including
some pretty old ones, refer to sending HUP signal as standard
way of restarting/resetting daemons.
Please tell the software authors about it,
On Tue, Oct 24, 2000 at 02:56:07PM -0700, Jordan Hubbard wrote:
So, who wants to do a proof-of-concept implementation for -current
which integrates with our existing rc.conf mechanism?
I was going to if no one else did.
Who ever does it should coordinate with Luke M @ NetBSD. He is willing
On Wed, Oct 25, 2000 at 08:14:01PM +0200, Gerhard Sittig wrote:
but I don't see FreeBSD having this level of "rc lib" as NetBSD
has in rc.subr
We would import the NetBSD rc.subr.
--
-- David ([EMAIL PROTECTED])
GNU is Not Unix / Linux Is Not UniX
To Unsubscribe: send mail to
On Tue, Oct 24, 2000 at 02:58:08PM -0700, Jordan Hubbard wrote:
With the NetBSD stuff, this is not immediately obvious though I guess
one could have a top level rc file with an explicit ordering similar to
our various subdir Makefiles,
Nope. All the /etc/rc.d/ files are scanned by `rcorder'.
Grrr !@#$^ Reply-To:...
On Wed, 25 Oct 2000 13:01:04 -0700, "David O'Brien" [EMAIL PROTECTED] said:
Nope. All the /etc/rc.d/ files are scanned by `rcorder'. `rcorder' then
creates a dependacy graph from information in each /etc/rc.d/ file. A
walk of the graph is done to output the
On Wed, Oct 25, 2000 at 09:42:23AM +0200, Andrea Campi wrote:
Maybe we could have a script to do the dependency check and "compile"
everything in a single big file?
Luke already has this support in NetBSD 1.5 for those who demand it, but
its a secret. ;-)
--
-- David ([EMAIL PROTECTED])
On Wed, Oct 25, 2000 at 04:04:13PM -0400, Garrett Wollman wrote:
Hmmm. We already have a program (called `tsort') which does this
(i.e., a topological sort). Does `rcorder' call `tsort' or does it
reinvent the wheel?
UTSL
lynx ftp://ftp.netbsd.org/pub/NetBSD-current/src/sbin/rcorder/
To
Gerhard Sittig writes:
What's new is:
- include the general config at the start (and yes, in every
single script -- but this should be neglectable in terms of
speed penalty and makes them work separately, too -- which is a
real big gain!)
This isn't really new; it's been nagging me
Grrr !@#%$^ Reply-To: header
On Wed, 25 Oct 2000 13:13:53 -0700, "David O'Brien" [EMAIL PROTECTED] said:
(i.e., a topological sort). Does `rcorder' call `tsort' or does it
reinvent the wheel?
UTSL
You could have simply answered the question. For the benefit of
everyone else: yes,
I was going to if no one else did.
Who ever does it should coordinate with Luke M @ NetBSD. He is willing
to make tweaks such that we could use as much of the NetBSD bits as
possible. He really hopes we [BSD] can standardize on this interface.
Well, it sounds like David is already
(i.e., a topological sort). Does `rcorder' call `tsort' or does it
reinvent the wheel?
UTSL
You could have simply answered the question. For the benefit of
everyone else: yes, it reinvents the wheel.
I personally don't have a problem with this; tsort should be
a library routine
On Wed, Oct 25, 2000 at 06:04:43AM +0700, Alexey Dokuchaev wrote:
On Tue, 24 Oct 2000, David O'Brien wrote:
On Tue, Oct 24, 2000 at 04:23:40PM +0700, Alexey Dokuchaev wrote:
Why can't I simply write kill -1 `cat /var/run/sendmail.pid`?
What about deamons that don't understand `kill
Garrett Rooney writes:
On Tue, Oct 24, 2000 at 04:49:40AM +0700, Alexey Dokuchaev wrote:
Well, would not be this stepping aside from BSD startup sequence, which we
all know and love? Having dozens of small files instead of pair of
big ones always frustrates me when I have to work with
One of the reasons for the numbers in the SysVR4 arena is to
set the order of execution so programs which other depend upon
are executed first. How does the NetBSD solve this problem.
Very coolly. The main rc script runs a script named `rcorder' to
generate the proper order.
I like the concept of them quite a bit. I think it definitely shows
some thought on how to keep the advantages of each system. I would
support a move toward a system like this. One thing that would be nice
is a database somewhere of which of services from /etc/rc.d are running.
I think
Well, would not be this stepping aside from BSD startup sequence, which we
all know and love? Having dozens of small files instead of pair of
big ones always frustrates me when I have to work with linux.
Install a binary package that needs to be started when the
system is booted and needs to
On Tue, 24 Oct 2000, Terry Lambert wrote:
Well, would not be this stepping aside from BSD startup sequence, which we
all know and love? Having dozens of small files instead of pair of
big ones always frustrates me when I have to work with linux.
Install a binary package that needs to
Well, we *already* have over a dozen /etc/rc.* files on -current. And
we *don't* have the advantage of a consistent interface to control all
the functions in /etc/rc. If you break things up, then if you need to
restart the mail server, just go "/etc/rc.d/sendmail restart". dhcpd?
Alexey Dokuchaev writes:
Well, we *already* have over a dozen /etc/rc.* files on -current. And
we *don't* have the advantage of a consistent interface to control all
the functions in /etc/rc. If you break things up, then if you need to
restart the mail server, just go "/etc/rc.d/sendmail
On Tue, Oct 24, 2000 at 08:59:36AM +, Terry Lambert wrote:
Oh... and the PROVIDE/REQUIRE/WANT lists really, really want to
be "per service name" rather than per program name, so I could,
for example, have a service that depends on "smtpserv", and not
care if it was sendmail or qmail or
The solution is very simple. Put a statically linked Perl in /sbin,
and write the startup system in Perl. For user convenience, it should
have a Gnome interface and a PostgreSQL backend, so we should also
put X and pgsql in /sbin.
--
Christopher Masto Senior Network Monkey
/me hands Chris SARCASM and /SARCASM
DocWilco
At 13:50 24-10-2000 -0400, you wrote:
The solution is very simple. Put a statically linked Perl in /sbin,
and write the startup system in Perl. For user convenience, it should
have a Gnome interface and a PostgreSQL backend, so we should
In message [EMAIL PROTECTED] "David O'Brien" writes:
: At BSDcon Luke M showed me what the NetBSD 1.5 rc files look like.
: They've moved them all to /etc/rc.d/ and made them very granular (as
: SVR4, but w/o leading numbers in the filenames). The NetBSD
: implementation also solved all the
In message [EMAIL PROTECTED] Bill Vermillion writes:
: One of the reasons for the numbers in the SysVR4 arena is to
: set the order of execution so programs which other depend upon
: are executed first. How does the NetBSD solve this problem.
The scripts themselves have the ordering
In message [EMAIL PROTECTED] "Matthew N. Dodd"
writes:
: On Mon, 23 Oct 2000, Brandon D. Valentine wrote:
: Well if you're that stubborn there's no reason that the "new" layout
: could not be compiled into a monolithic script. In fact perhaps you
: could be the one to step forward and write
In message [EMAIL PROTECTED] "Jacques A. Vidrine" writes:
: By the way, the author of this stuff (Luke Mewburn) says he'll post a
: summary of the design and implementation issues to this list in a few
: days.
I talked to Luke at BSDcon about many issues. He's very keen on
increasing the
On Tue, Oct 24, 2000 at 16:14 +0700, Alexey Dokuchaev wrote:
On Tue, 24 Oct 2000, Terry Lambert wrote:
Well, would not be this stepping aside from BSD startup
sequence, which we all know and love? Having dozens of
small files instead of pair of big ones always frustrates
me when I
On Tue, Oct 24, 2000 at 05:05:49AM +0700, Alexey Dokuchaev wrote:
Still, it would be better if I could choose between "classical" and "new"
startup layout, say, somewhere at the installation stage.
Supporting two very different schemes is a support nightmare. And
giveing good test coverage
On Tue, Oct 24, 2000 at 04:49:40AM +0700, Alexey Dokuchaev wrote:
Having dozens of small files instead of pair of big ones always
frustrates me when I have to work with linux.
Maybe, but the greatly increased functionality makes it worth it.
--
-- David ([EMAIL PROTECTED])
GNU is
On Tue, Oct 24, 2000 at 04:23:40PM +0700, Alexey Dokuchaev wrote:
Why can't I simply write kill -1 `cat /var/run/sendmail.pid`?
What about deamons that don't understand `kill -HUP'? Sendmail didn't
until very reciently. ``/etc/rc.d/some-deamon restart'' does the right
thing reguardless how
[redirected to just -current; I'm not sure what this has to do with -net]
I agree. I've been using them for a while on my dog slow Windows CE
machine. There were some minor issues when they were first committed
to NetBSD on some platforms (due to a too early use of ps and some
brokeness in
The scripts themselves have the ordering dependencies. The startup
system runs them in the proper order. I don't know if this is
pre-computed or redone each boot.
I'm really curious about this, myself. One of the reasons the SYSV
scripts have the numeric prefix is so that you know exactly
On Tue, Oct 24, 2000 at 02:58:08PM -0700, Jordan Hubbard wrote:
The scripts themselves have the ordering dependencies. The startup
system runs them in the proper order. I don't know if this is
pre-computed or redone each boot.
I'm really curious about this, myself. One of the reasons
On Tue, Oct 24, 2000 at 02:58:08PM -0700, Jordan Hubbard wrote:
The scripts themselves have the ordering dependencies. The startup
system runs them in the proper order. I don't know if this is
pre-computed or redone each boot.
I'm really curious about this, myself. One of the reasons
On Tue, Oct 24, 2000 at 02:58:08PM -0700, Jordan Hubbard thus spoke:
The scripts themselves have the ordering dependencies. The startup
system runs them in the proper order. I don't know if this is
pre-computed or redone each boot.
I'm really curious about this, myself. One of the
Jordan Hubbard writes:
[redirected to just -current; I'm not sure what this has to do with -net]
I agree. I've been using them for a while on my dog slow Windows CE
machine. There were some minor issues when they were first committed
to NetBSD on some platforms (due to a too early use of
So, who wants to do a proof-of-concept implementation for -current
which integrates with our existing rc.conf mechanism? In order to
obey POLA, we should at least have the separate scripts switch off the
same knobs whenever possible.
It's something I'd be willing to do, I guess. I have some
On Tue, 24 Oct 2000, David O'Brien wrote:
On Tue, Oct 24, 2000 at 04:23:40PM +0700, Alexey Dokuchaev wrote:
Why can't I simply write kill -1 `cat /var/run/sendmail.pid`?
What about deamons that don't understand `kill -HUP'? Sendmail didn't
until very reciently. ``/etc/rc.d/some-deamon
and, to reply a second time to this message, it is recomputed at each
boot... the rc and rc.shutdown scripts both run rcorder to do it, with
rc.shutdown reversing the order.
Ah, OK, sorry - I must have missed this the first time around.
I'll have to investigate the workings of rcorder then.
I'm in the midst of trying to install NetBSD so I can look at this. If
no one else steps forward to do it, I can put together a patch.
I've had several replies, so why don't we all look into this a bit and
see which one of us actually manages to have enough steam to do it
after the analysis
At 2:58 PM -0700 10/24/00, Jordan Hubbard wrote:
The scripts themselves have the ordering dependencies. The
startup system runs them in the proper order. I don't know
if this is pre-computed or redone each boot.
I'm really curious about this, myself. One of the reasons the SYSV
scripts
On Tue, Oct 24, 2000 at 11:04:55PM -0400, Garance A Drosihn wrote:
One should have some other script that you could run, which
would look thru all the rc files and just list which order
they will be run at startup (or at shutdown). That way you
could find out the order for a given set of
On Mon, Oct 23, 2000 at 01:05:27AM -0700, David O'Brien wrote:
On Sun, Oct 22, 2000 at 09:41:51PM -0400, Bill Vermillion wrote:
One of the reasons for the numbers in the SysVR4 arena is to
set the order of execution so programs which other depend upon
are executed first. How does the
On Mon, 23 Oct 2000, Brian O'Shea wrote:
Sounds interesting. To add a new rc script to the system, do you have
to add an entry to an "rc order list" somewhere (in addition to adding
the new script)? How is that handled? The nice (or clumsy, depending
on your point of view) part about the SysV
On Mon, Oct 23, 2000 at 05:07:42PM -0400, Brandon D. Valentine wrote:
On Mon, 23 Oct 2000, Brian O'Shea wrote:
Sounds interesting. To add a new rc script to the system, do you have
to add an entry to an "rc order list" somewhere (in addition to adding
the new script)? How is that handled?
On Mon, Oct 23, 2000 at 05:07:42PM -0400, Brandon D. Valentine wrote:
Hmm I don't have any NetBSD machines running the later 1.5 revisions
yet, so I've not seen the new scripts,
lynx ftp://ftp.netbsd.org/pub/NetBSD-current/src/etc/rc.d/
To Unsubscribe: send mail to [EMAIL PROTECTED]
with
On Mon, 23 Oct 2000, David O'Brien wrote:
On Mon, Oct 23, 2000 at 05:07:42PM -0400, Brandon D. Valentine wrote:
Hmm I don't have any NetBSD machines running the later 1.5 revisions
yet, so I've not seen the new scripts,
lynx ftp://ftp.netbsd.org/pub/NetBSD-current/src/etc/rc.d/
Thanks, I was
On Mon, Oct 23, 2000 at 02:25:40PM -0700, David O'Brien wrote:
On Mon, Oct 23, 2000 at 05:07:42PM -0400, Brandon D. Valentine wrote:
Hmm I don't have any NetBSD machines running the later 1.5 revisions
yet, so I've not seen the new scripts,
lynx
On Sun, 22 Oct 2000, David O'Brien wrote:
On Sat, Oct 21, 2000 at 11:05:37AM -0700, Jordan Hubbard wrote:
I wish to update rc.network6 and introduce rc.firewall6.
H. I must confess that I see /etc as getting rather cluttered
these days. Is there no way to perhaps collapse some
On Tue, Oct 24, 2000 at 04:49:40AM +0700, Alexey Dokuchaev wrote:
Well, would not be this stepping aside from BSD startup sequence, which we
all know and love? Having dozens of small files instead of pair of
big ones always frustrates me when I have to work with linux.
well, it's a single
On Mon, 23 Oct 2000, Garrett Rooney wrote:
On Tue, Oct 24, 2000 at 04:49:40AM +0700, Alexey Dokuchaev wrote:
Well, would not be this stepping aside from BSD startup sequence, which we
all know and love? Having dozens of small files instead of pair of
big ones always frustrates me when I
On Tue, 24 Oct 2000, Alexey Dokuchaev wrote:
Still, it would be better if I could choose between "classical" and "new"
startup layout, say, somewhere at the installation stage.
Well if you're that stubborn there's no reason that the "new" layout
could not be compiled into a monolithic script.
On Mon, 23 Oct 2000, Brandon D. Valentine wrote:
On Tue, 24 Oct 2000, Alexey Dokuchaev wrote:
Still, it would be better if I could choose between "classical" and "new"
startup layout, say, somewhere at the installation stage.
Well if you're that stubborn there's no reason that the "new"
On Tue, Oct 24, 2000 at 05:26:07AM +0700, Alexey Dokuchaev wrote:
On Mon, 23 Oct 2000, Brandon D. Valentine wrote:
On Tue, 24 Oct 2000, Alexey Dokuchaev wrote:
Still, it would be better if I could choose between "classical" and "new"
startup layout, say, somewhere at the installation
On Mon, 23 Oct 2000, Garrett Rooney wrote:
That's an idea! Gotta co recent -CURRENT right now!
might want to port the netbsd code first, since AFAIK this stuff isn't
in current ;-)
Indeed it's not, but nice to seem him so eager. =)
--
Brandon D. Valentine [EMAIL PROTECTED]
"Few things are
On Mon, 23 Oct 2000, Brandon D. Valentine wrote:
Well if you're that stubborn there's no reason that the "new" layout
could not be compiled into a monolithic script. In fact perhaps you
could be the one to step forward and write the code to compile that
script. ;-)
Indeed, given the
On Mon, Oct 23, 2000 at 02:25:40PM -0700, David O'Brien wrote:
On Mon, Oct 23, 2000 at 05:07:42PM -0400, Brandon D. Valentine wrote:
Hmm I don't have any NetBSD machines running the later 1.5 revisions
yet, so I've not seen the new scripts,
lynx
On Sun, Oct 22, 2000 at 03:39:57PM -0700, David O'Brien thus spoke:
On Sat, Oct 21, 2000 at 11:05:37AM -0700, Jordan Hubbard wrote:
I wish to update rc.network6 and introduce rc.firewall6.
H. I must confess that I see /etc as getting rather cluttered
these days. Is there no way to
I wish to update rc.network6 and introduce rc.firewall6.
H. I must confess that I see /etc as getting rather cluttered
these days. Is there no way to perhaps collapse some of the most
related functionality into single files and start passing arguments
or something? Just a comment..
-
-On [20001021 20:10], Jordan Hubbard ([EMAIL PROTECTED]) wrote:
I wish to update rc.network6 and introduce rc.firewall6.
H. I must confess that I see /etc as getting rather cluttered
these days. Is there no way to perhaps collapse some of the most
related functionality into single files
However, Umemoto-san and me will discuss this, since we [he mostly] have
been working on this for the last few months.
Sounds good to me. My comments were, just to make it clear again,
just food for thought and not out-and-out objections. If even 47 more
files in /etc is what it takes to get
70 matches
Mail list logo
