[Pkg-sysvinit-devel] Early RW filesystem: PROPOSAL AND VOTE REQUEST

[Henrique de Moraes Holschuh]

On Sat, 16 Sep 2006, Henrique de Moraes Holschuh wrote:
> Well, I think we will need to ask a TC rulling for this one, given that
> people could not decide on a name for it.

And the thread in -devel shows I am quite right, again.  I am in no mood to
wait for it to settle, as past experience says it will not help any to wait.
I hereby propose that:

1. CONSIDERING THAT our highest priority is to have an early writeable
filesystem that is technically sound;

2. CONSIDERING THAT whether /var/run or /var/lock is in a tmpfs or not is
not a high priority issue, as long as it is not tied to the early writeable
filesystem;

3. CONSIDERING THAT using any mountpoint that could be shadowed by a later
mount requires kernel 2.6 features, but that such techniques are only needed
for the early writeable filesystem if our choice of mountpoint is not inside
a part of the filesystem that is already required to be part of the root
partition;

4. IN LIGHT OF the ammount of discursion this whole issue is causing in
debian-devel, mostly because of the usage of tmpfs in /var/run and
/var/lock;

I hereby propose that:

1. We use /lib/init/rw as the mountpoint of the early-writeable filesystem,
thus decoupling it from any other tangencial issues;

2. We use a tmpfs limited to 5MB in size, options exec,dev,nosuid, ownership
root:root, mode 01755 for the early-writeable filesystem in Linux systems
(and defer the correct implementation for other kernels to their porter
teams, but most of them have something akin to tmpfs);

3. We request a Technical Comitee ruling about the default usage of tmpfs
for /var/run and /var/lock and their default sizes, and also whether
packages are required to support ephemeral /var/run and /var/lock
filesystems regardless of the choice of default;

4. We revert the /var/run and /var/lock tmpfs changes while we wait for the
TC decision;

5. We investigate the difficulty and impact of supporting an environment
variable with the location of the early-writeable filesystem, and deploy it
if it is not too hard, otherwise we proceed without such a variable;

Please reply with your thoughts, and whether you agree with this or not.
Let's get this over quickly, if at all possible...

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 481 bytes
Desc: Digital signature
Url : http://lists.alioth.debian.org/pipermail/pkg-sysvinit-devel/attachments/20060918/9f3fdcde/attachment.pgp