| 1 |
On Sat, Feb 29, 2020 at 4:33 AM Wols Lists <antlists@××××××××××××.uk> wrote: |
| 2 |
> |
| 3 |
> I just have a massive swap space, and /var/tmp/portage is a tmpfs. So |
| 4 |
> everything gets a fast tmpfs build, and it spills into swap as required |
| 5 |
> (hopefully almost never). |
| 6 |
> |
| 7 |
|
| 8 |
I can articulate a bunch of reasons that on paper say that this is the |
| 9 |
best approach. |
| 10 |
|
| 11 |
In practice I've found that swap on linux is sub-optimal at best. I |
| 12 |
only enable swap when I absolutely have to as a result. I'll reduce |
| 13 |
-j to lower memory demand before adding swap usually. On more |
| 14 |
RAM-constrained hosts I'll enable swap when building specific |
| 15 |
packages, or try to avoid those packages entirely. |
| 16 |
|
| 17 |
Maybe something has changed in the last few years and swap is actually |
| 18 |
useful, but I'm skeptical. I always tend to end up with GB of free |
| 19 |
RAM and a churning hard drive when I enable it. On SSD I'm sure it |
| 20 |
will perform better, but then I'm running through erase cycles |
| 21 |
instead. |
| 22 |
|
| 23 |
Like I said, on paper adding swap should only make things better. |
| 24 |
But, that is only true if the kernel makes the correct choices about |
| 25 |
prioritizing swap vs cache use. Sure, I could set swappiness to zero |
| 26 |
or whatever, but then that just turns swap into a NOOP best case and |
| 27 |
it isn't like I have OOM issues, so why bother?. |
| 28 |
|
| 29 |
-- |
| 30 |
Rich |
Archives