Re: [gentoo-user] Running out of space on /var partition

[Michael Mol]

On Wed, Jul 27, 2011 at 10:40 AM, Neil Bothwick <neil@××××××××××.uk> wrote:
> On Wed, 27 Jul 2011 14:41:33 +0100, Peter Humphrey wrote:
>
>> Doesn't do that here. When tmpfs is full it starts being swapped out to
>> the swap partition. Perhaps you didn't have any swap at the time.
>
> The default size for a tmpfs filesystem is half the physical RAM, unless
> you specify more as a mount option, it will never use significant
> amounts of swap.
>
> I wonder how effective tmpfs is for PORTAGE_TMPDIR as the builds that
> need a lot of disk space can often require a fair bit of memory too, and
> tmpfs is using it all.

portage (by default) cleans up after itself after each ebuild, so
there's no leakage across ebuilds, but yeah; if you don't have enough
RAM for it, stuffing your files and your active processes in the RAM
won't work. That's a pretty simple concept.

tmpfs for PORTAGE_TMPDIR is great, as long as you have the RAM for it.
When I first started using Gentoo, I did. Lately, builds have gotten
large enough (especially when I added -ggdb to CFLAGS) that I had to
stop using it.

However, I can see an argument for tmpfs to still be useful, even if
you don't have enough RAM for it, but you do have swap space. If you
allow tmpfs to be backed by swap, you can avoid the (unnecessary in
this case) bookkeeping done by normal disk filesystems like
ext{2|3|4}. So long as your swap partition is as fast as the partition
your erstwhile disk filesystem sits on, it'd be a theoretical gain.
(Though if you're using a swap *file* as opposed to a swap partition,
you'd lose that gain. So YMMV.)

With that in mind, I think I'll re-enable tmpfs when I get home, and
set it to act as a 16GB filesystem. I only have 6GB of RAM, but tmpfs
won't consume more RAM than it has data to fill pages, and the use of
swap shouldn't be terrible.

My understanding of tmpfs is that it's implemented as a thin layer on
top of the file page cache, and that as active processes need more
RAM, the file page cache is the first thing to be sacrificed, being
flushed to disk. (Presumably, in the case of tmpfs, 'disk' would be
'swap') So truly active process memory should have no more trouble
remaining in physical RAM than if tmpfs weren't there; if tmpfs
weren't there, the data would still sit in the file cache, but would
have to pass through a filesystem like ext{2|3|4} (or whatever) on its
way to disk.

-- 
:wq