This one time, at band camp, Blars Blarson wrote:
In article [EMAIL PROTECTED]
[EMAIL PROTECTED] writes:
I won't debate whether this is true in general, bug it is certainly
unnecessary in the case of pump. I have specifically added code to
deal with the inability to write to /var/run by making
This one time, at band camp, Marco d'Itri wrote:
On Apr 07, Steve Langasek [EMAIL PROTECTED] wrote:
Duh. Did you miss the part where people were talking about *amending the
FHS because the FHS is flawed*?
Yes: I did not agree with them.
And looks like I was right, because as I showed nearly
This one time, at band camp, Marco d'Itri wrote:
On Apr 07, Thomas Hood [EMAIL PROTECTED] wrote:
* pam, shadow
Allow either /etc/nologin or /run/nologin to prevent non-root logins
Use a symlink.
* util-linux
Use /run/mtab for mount's statefile
Use a symlink.
A symlink
This one time, at band camp, Thomas Hood wrote:
On Sat, 2003-04-12 at 10:12, Anthony DeRobertis wrote:
On Wed, 2003-04-09 at 14:17, Thomas Hood wrote:
* ppp
* Change /usr/sbin/pppd to:
* Store PID in /run/, not in /var/run/
Why? Is the goal to make PPP-mounter /var to work?!
I
This one time, at band camp, Thomas Hood wrote:
(Re: /etc/nologin)
A dangling symlink should be considered like a missing file.
Yes, that would work. However, having separate /etc/nologin
and /run/nologin looks like a useful feature, as I mentioned
earlier.
For clarification, (and this is
This one time, at band camp, Anthony DeRobertis wrote:
On Wed, 2003-04-09 at 14:17, Thomas Hood wrote:
The proposed new directory is for files similar to those in /var/run/
that are not just variable and unshareable but also local -- i.e., they
must be writable independently of network
This one time, at band camp, Thomas Hood wrote:
* base-files
Add /run/ directory
#191036: create /run for programs that run before /var is mounted
* pam, shadow
Allow either /etc/nologin or /run/nologin to prevent nonroot login
#191037: Allow both /etc/nologin and /run/nologin
Jamie Wilkinson wrote:
That is right! The core of the matter is not whether
filesystems need to be mounted over the network or not,
but whether the parts of the filesystem you are attempting
to write to are on the root filesystem or not.
The essence of /run/ had better not include that it be
This one time, at band camp, Marco d'Itri wrote:
/etc has a static nature. See the note on /etc/mtab under Table
3.7.3.1. It is also for configuration files. /etc/adjtime is neither.
Then propose to change FHS.
Welcome to the discussion.
--
[EMAIL PROTECTED]
This one time, at band camp, Marco d'Itri wrote:
On Apr 07, Thomas Hood [EMAIL PROTECTED] wrote:
(I forgot to mention) J.W.'s patch does more than what a mere symlink
would do. By making programs sensitive both to /run/nologin and
/etc/nologin, it becomes possible for the administrator to
This one time, at band camp, Thomas Hood wrote:
Jamie Wilkinson wrote:
That is right! The core of the matter is not whether
filesystems need to be mounted over the network or not,
but whether the parts of the filesystem you are attempting
to write to are on the root filesystem or not.
The
On Mon, Apr 28, 2003 at 06:50:15PM +1000, Jamie Wilkinson wrote:
This one time, at band camp, Thomas Hood wrote:
Jamie Wilkinson wrote:
That is right! The core of the matter is not whether
filesystems need to be mounted over the network or not,
but whether the parts of the filesystem you
Sorry to reopen this at such a late date, but I'm way behind on -devel.
Hi, I'm Karl and I maintain login and passwd.
Thomas Hood [EMAIL PROTECTED] writes:
* pam, shadow
Allow either /etc/nologin or /run/nologin to prevent non-root logins
I don't like the idea of having multiple files
On Mon, 2003-04-28 at 18:30, [EMAIL PROTECTED] wrote:
I don't like the idea of having multiple files to turn off logins. (I
can't log into my system, and /etc/nologin doesn't exist! What? didn't you
know about this *other* file?) I also don't want to solve this with a
symlink.
Yes, let's
On Mon, Apr 28, 2003 at 11:35:58PM +0200, Thomas Hood wrote:
On Mon, 2003-04-28 at 18:30, [EMAIL PROTECTED] wrote:
It would also be nice to have some blessing of /run in the policy first,
but that doesn't seem terribly likely.
What is more important for now is whether there is broad enough
This one time, at band camp, Steve Langasek wrote:
Shell pseudocode was posted to this thread (a month ago?) that showed
how the init scripts could handle the requirement that /run be mounted
early, even if it's not on the root fs. The init scripts already
include special handling of /proc and /,
This one time, at band camp, [EMAIL PROTECTED] wrote:
Sorry to reopen this at such a late date, but I'm way behind on -devel.
Hi, I'm Karl and I maintain login and passwd.
Thomas Hood [EMAIL PROTECTED] writes:
* pam, shadow
Allow either /etc/nologin or /run/nologin to prevent non-root
On Tue, 2003-04-15 at 15:19, Thomas Hood wrote:
Unfortunately you seem to be wrong, at least with regard to
bind version 1:8.3.4-4.
Ah. That'd explain it. I'm using bind9.
signature.asc
Description: This is a digitally signed message part
On 8 April 2003 Marco d'Itri [EMAIL PROTECTED] wrote:
On Apr 07, Thomas Hood [EMAIL PROTECTED] wrote:
A difficulty is that only a whole options { ... };
statement can be included from the named configuration file,
not just the forwarders { ... }; statement inside it.
You can include a file even
On Thu, Apr 10, 2003 at 07:27:42PM -0400, Jeremy Jackson wrote:
Overall I think it is wonderful to see support for read-only root being
worked on.
On Wed, 2003-04-09 at 15:41, Matthew Garrett wrote:
Jeremy Jackson wrote:
(doing this with bind mounts)
2.2 kernels are out though.
In article [EMAIL PROTECTED]
[EMAIL PROTECTED] writes:
I won't debate whether this is true in general, bug it is certainly
unnecessary in the case of pump. I have specifically added code to
deal with the inability to write to /var/run by making pump fall back
to using TCP sockets.
It will also
Blars Blarson [EMAIL PROTECTED] wrote:
It will also need to cope with writing to /var/run on the root
partition, having /var mounted, and later processes not being able to
open the file since the /var/run directory on the root disk is
inaccessable.
It does. If the Unix-domain socket does
On Sat, 2003-04-12 at 23:56, Herbert Xu wrote:
* Change /sbin/pump to:
* Store PID in /run, not in /var/run
I won't debate whether this is true in general, bug it is certainly
unnecessary in the case of pump. I have specifically added code to
deal with the inability to write to
On Sat, 2003-04-12 at 10:12, Anthony DeRobertis wrote:
On Wed, 2003-04-09 at 14:17, Thomas Hood wrote:
* ppp
* Change /usr/sbin/pppd to:
* Store PID in /run/, not in /var/run/
Why? Is the goal to make PPP-mounter /var to work?!
I suppose someone might want to mount /var/
Thomas Hood [EMAIL PROTECTED] wrote:
Unnecessary; but would using /run for the pidfile be a better
(e.g., simpler) solution?
If not then do you think the TCP-socket approach is the way
to deal with every program that writes a pidfile when /var/
may be absent?
Pump doesn't write pid
On Wed, 2003-04-09 at 10:13, Marco d'Itri wrote:
Leaving /etc/adjtime as is and telling admins to move it and use a
symlink is a FHS violation because /etc/adjtime is.
What part of FHS does not apply to local changes you did not
understand?
No part. Please read my message again. The
On Wed, 2003-04-09 at 14:17, Thomas Hood wrote:
* ppp
* Change /usr/sbin/pppd to:
* Store PID in /run/, not in /var/run/
Why? Is the goal to make PPP-mounter /var to work?! If so, pppd has to
be moved to /sbin.
* pump
* Add /etc/pump directory
* Change /sbin/pump to:
Anthony DeRobertis [EMAIL PROTECTED] wrote:
* pump
* Add /etc/pump directory
* Change /sbin/pump to:
* Store PID in /run, not in /var/run
Quite. Programs in /sbin shouldn't, in general, be using /var/run.
I won't debate whether this is true in general, bug it is certainly
* Jeremy Jackson [EMAIL PROTECTED] [030410 23:59]:
Can someone point me the message(s) discussing /run (and why not
/etc/run) - I would like to think that adding another directory off /
should be avoided. /etc/run sounds nice, unless you want to support
booting before /etc is mounted...
On
(I am resending this because the earlier version contained
a few minor errors. I suppose I should put this on the web
somewhere and post the link.)
(I repeat the call for people to look for files that are in
/etc/ but shouldn't be.)
Here are:
* an updated list of wishes filed,
* an updated TODO
On Wed, 2003-04-09 at 20:22, Jeremy Jackson wrote:
Clearly what is needed here is an API for resolver updates
[...]
What you describe is roughly what I wrote up in my last TODO
message.
In Debian, there should possibly be a policy decided upon.
(what the dir is, what API is, etc)
I don't
Can someone point me the message(s) discussing /run (and why not
/etc/run) - I would like to think that adding another directory off /
should be avoided. /etc/run sounds nice, unless you want to support
booting before /etc is mounted...
Cheers,
Jeremy
On Wed, 2003-04-09 at 14:17, Thomas Hood
Overall I think it is wonderful to see support for read-only root being
worked on.
On Wed, 2003-04-09 at 15:41, Matthew Garrett wrote:
Jeremy Jackson wrote:
(doing this with bind mounts)
2.2 kernels are out though.
As are BSDs. I have no idea whether the Hurd supports bind mounting.
33 matches
Mail list logo