| 1 |
Daniel Robbins wrote: |
| 2 |
|
| 3 |
> But right now I am getting held up by a kernel bug. I can mount a |
| 4 |
> loopback fs, and try to copy files to it. I get a "no space left on |
| 5 |
> device." I can then repeat the exact same cp command, and a few seconds |
| 6 |
> later it works. Seems to be some weird synchronization bug in the |
| 7 |
> kernel. I'm using gentoo-dev-sources-2.6.0 and I'm wondering if anyone |
| 8 |
> has experienced anything similar. This is on amd64 but the loopback code |
| 9 |
> in the kernel should be arch-independent. |
| 10 |
|
| 11 |
Might help to prune /dev a little. I spent a good while wiping out the |
| 12 |
contents of /dev after creating generic-mips entries for the mips |
| 13 |
netboot. Things like six different pty*, ttyS* (or tty* in general), |
| 14 |
and other things like midi entries, etc.. really aren't needed (unless |
| 15 |
anyone plans on composing some midi tunes on a livecd while the install |
| 16 |
goes forth). Cut about /dev 1000+ entries down to around ~250 (or a |
| 17 |
close approximation of such). |
| 18 |
|
| 19 |
That might help skirt around the lack of space issue some. All in all, |
| 20 |
I created an ext2 fs w/ only 1500 inodes for the netboot and had no |
| 21 |
problems, although for a livecd, I know this will not be sufficient. |
| 22 |
|
| 23 |
I also used 4k blocksizes in the netboot. It seemed to have allowed the |
| 24 |
filesystem to take up a little bit less space than using the standard 1k |
| 25 |
block size (saved 200-300KB), although I don't know if this would have |
| 26 |
any impact either on the problem your having. It was mainly and experiment. |
| 27 |
|
| 28 |
|
| 29 |
--Kumba |
| 30 |
|
| 31 |
-- |
| 32 |
"Such is oft the course of deeds that move the wheels of the world: |
| 33 |
small hands do them because they must, while the eyes of the great are |
| 34 |
elsewhere." --Elrond |
| 35 |
|
| 36 |
|
| 37 |
-- |
| 38 |
gentoo-releng@g.o mailing list |
Archives