| 1 |
On Wed, 7 Jan 2015 15:22:46 -0500, Rich Freeman wrote: |
| 2 |
|
| 3 |
> > That's not what I meant, but I see your point about using --reflink if |
| 4 |
> > making copies. My thought was to forget the whole tmpfs and copying |
| 5 |
> > think, set KEEP WORK in FEATURES and use XFS for PORTAGE_TMPDIR. That |
| 6 |
> > way most of the work is taking place in the cache without the |
| 7 |
> > frequent disk writes of other filesystems. I would expect XFS to be |
| 8 |
> > faster for this job, but have no data to support that assumption. |
| 9 |
> > |
| 10 |
> > Or you could just put TMPDIR on btrfs and snapshot after each emerge. |
| 11 |
> > |
| 12 |
> |
| 13 |
> If you're not going to move it but just leave it all in |
| 14 |
> TMPDIR forever, then by all means use whatever filesystem you like. |
| 15 |
|
| 16 |
Of course, but the original question was about improving performance (via |
| 17 |
tmpfs) and a heavily cached filesystem should go partway towards that. |
| 18 |
Mind you, ext2 would also help as it would avoid the journal overhead. |
| 19 |
|
| 20 |
> Sure, |
| 21 |
> you could snapshot it and clean it, but that might not be the |
| 22 |
> cleanest way. |
| 23 |
|
| 24 |
You'd have to manager the snapshots, but then you'd have to manage any |
| 25 |
form of copies if you don't want to run out of disk space, but at least |
| 26 |
it would be fast. |
| 27 |
|
| 28 |
|
| 29 |
-- |
| 30 |
Neil Bothwick |
| 31 |
|
| 32 |
Welcome to the world of Windows 95. Stay a while -- stay foooreveeer. |
Archives