| 1 |
On 19/10/2019 16:24, Mick wrote: |
| 2 |
> On Saturday, 19 October 2019 14:11:26 BSTmad.scientist.at.large@××××××××.com |
| 3 |
> wrote: |
| 4 |
>> Do systems run different memory management when swap is on versus no swap? |
| 5 |
|
| 6 |
> The answer to this question is an unqualified yes, although you do not define |
| 7 |
> your meaning of "different memory management". The existence of swap space |
| 8 |
> and the kernel's swappiness setting will change the way memory is dynamically |
| 9 |
> allocated to processes at runtime and may affect the responsiveness of your |
| 10 |
> system. |
| 11 |
|
| 12 |
Memory management was massively rewritten roundabout kernel 2.4. |
| 13 |
|
| 14 |
The original swap algorithm NEEDED twice ram as swap. And when Linus |
| 15 |
ripped out all the "optimisation", the vanilla kernels only needed to |
| 16 |
touch swap, and if they didn't have twice ram they would crash. |
| 17 |
|
| 18 |
At that point, the recommendation changed to "no swap is fine, twice or |
| 19 |
more is fine, just don't have swap less than twice ram". |
| 20 |
|
| 21 |
My personal rule is to take the motherboard's max ram, double it, and |
| 22 |
create a swap partition that size on every disk. So my current desktop |
| 23 |
system has 80GB of ram/swap - 4x4GB slots times 2 disk drives. And my |
| 24 |
new system has 4x8GB so that'll be 160GB!!! HOWEVER - Richard Brown of |
| 25 |
SUSE said that's dangerous - if somebody fork-bombs you it'll take a |
| 26 |
long time to fill that much swap and regaining control of your system |
| 27 |
could well be a big red switch job. |
| 28 |
|
| 29 |
Cheers, |
| 30 |
Wol |
Archives