1 |
On Sat, 26 May 2012 17:17:54 -0500, Dale wrote: |
2 |
|
3 |
> > It makes sure that /run is available and writeable early in the boot |
4 |
> > process, whereas /var/run may not be and / may be mounted ro. |
5 |
|
6 |
> Mine wouldn't be since I have /var on a separate partition. I guess the |
7 |
> devs are getting ready for the ultimate screwup udev and friends is |
8 |
> putting in place. |
9 |
|
10 |
No, it's avoiding a screwup. If you have /var on a separate partition, as |
11 |
I do, and something early in the boot process writes to /var/run |
12 |
(or /var/lock) whatever is written disappears when the var filesystem is |
13 |
mounted on /var. Using a tmpfs in / prevent this. |
14 |
|
15 |
The alternative is to require /var is on the same filesystem as / or |
16 |
mounted from an initramfs. ISTR you were rather against such a move. |
17 |
|
18 |
This move makes perfect sense, volatile but essential data is kept in ram |
19 |
rather than on a filesystem that may not always be available. If you are |
20 |
really bothered about the maximum size, remount it, although an option to |
21 |
specify this in rc.conf may possibly be useful in some situations. |
22 |
|
23 |
|
24 |
-- |
25 |
Neil Bothwick |
26 |
|
27 |
(A)bort, (R)etry, (P)retend this never happened... |