| 1 |
On Thu, Sep 1, 2016 at 8:41 AM, Michael Mol <mikemol@×××××.com> wrote: |
| 2 |
> |
| 3 |
> The defaults for vm.dirty_bytes and vm.dirty_background_bytes are, IMO, badly |
| 4 |
> broken and an insidious source of problems for both regular Linux users and |
| 5 |
> system administrators. |
| 6 |
> |
| 7 |
|
| 8 |
It depends on whether you tend to yank out drives without unmounting |
| 9 |
them, or if you have a poorly-implemented database that doesn't know |
| 10 |
about fsync and tries to implement transactions across multiple hosts. |
| 11 |
|
| 12 |
The flip side of all of this is that you can save-save-save in your |
| 13 |
applications and not sit there and watch your application wait for the |
| 14 |
USB drive to catch up. It also allows writes to be combined more |
| 15 |
efficiently (less of an issue for flash, but you probably can still |
| 16 |
avoid multiple rounds of overwriting data in place if multiple |
| 17 |
revisions come in succession, and metadata updating can be |
| 18 |
consolidated). |
| 19 |
|
| 20 |
For a desktop-oriented workflow I'd think that having nice big write |
| 21 |
buffers would greatly improve the user experience, as long as you hit |
| 22 |
that unmount button or pay attention to that flashing green light |
| 23 |
every time you yank a drive. |
| 24 |
|
| 25 |
-- |
| 26 |
Rich |
Archives