/var size from 8M to 50M on ALIX2 boards

[rick]

We hebben zat geheugen vrij (140M), een grotere /var (en hierbij dus ook /tmp) zorgt voor meer ruimte tijdens debuggen en statistieken.

[rick]

Thinking of killing the default link from /tmp -> /var/tmp allowing us to make two memory disks. This gives us the added benefit of making /var more robust and file-size-stable, as huge files in /tmp will not affect the /var system anymore. The default /tmp should be 5M and can grow as big as 50M on ALIX boards.

Adding more space to /var seems a bit useless.

[rick]

HybridRick2# df
Filesystem   1K-blocks   Used Avail Capacity  Mounted on
/dev/ada0s2a    403288 333100 37925    90%    /
devfs                1      1     0   100%    /dev
/dev/md0          4380   2536  1496    63%    /etc
/dev/md1          9116   1404  6984    17%    /var
/dev/md2         19740     64 18100     0%    /tmp

Dat lijkt er meer op.

[rick]

In 10828/hybrid:

Use an seperate /tmp var disk to provide an more robust system. A full /tmp is
not and issue while an full /var could cause all kind of weird and wonderfull
magic, as PID administration if for example done at /var/run, etc...

Done by undoing the stuff to make it happen by nanobsd.sh, as their is no
direct flag to control it.

As late_customize is an hook never used, make sure it does not break our
package build sub-system.

Related-To: nodefactory:ticket:162

[rick]

Created an seperate /tmp instead, defaults to 20M in size.