| 1 |
On Tue, Jan 19, 2016 at 10:56:21PM +0100, lee wrote: |
| 2 |
> Alec Ten Harmsel <alec@××××××××××××××.com> writes: |
| 3 |
> > |
| 4 |
> > Depends on how the load is. Right now I have a 500GB HDD at work. I use |
| 5 |
> > VirtualBox and vagrant for testing various software. Every VM in |
| 6 |
> > VirtualBox gets a 50GB hard disk, and I generally have 7 or 8 at a time. |
| 7 |
> > Add in all the other stuff on my system, which includes a 200GB dataset, |
| 8 |
> > and the disk is overcommitted. Of course, none of the VirtualBox disks |
| 9 |
> > use anywhere near 50GB. |
| 10 |
> |
| 11 |
> True, that's for testing when you do know that the disk space will not |
| 12 |
> be used and have no trouble when it is. When you have the VMs in |
| 13 |
> production and users (employees) using them, you don't know when they |
| 14 |
> will run out of disk space and trouble ensues. |
| 15 |
|
| 16 |
Almost. Here is an equal example: I am an admin on an HPC cluster. We |
| 17 |
have a shared Lustre filesystem that people store work files in while |
| 18 |
they are running jobs. It has around 1PB of capacity. As strange as this |
| 19 |
may sound, this filesystem is overcommitted (we have 20,000 cores, |
| 20 |
that's only 52GB per core, not even close to enough for more than half a |
| 21 |
year of data accumulation). Unused data is deleted after 90 days, which |
| 22 |
is why it can be overcommitted. |
| 23 |
|
| 24 |
Extending this to a more realistic example without automatic data |
| 25 |
deletion is trivial. Imagine you are a web hosting provider. You allow |
| 26 |
each client unlimited disk space, so you're automatically overcommitted. |
| 27 |
In the aggregate, even though one client may increase their usage |
| 28 |
extremely quickly, total usage rises slowly, giving you more than enough |
| 29 |
time to increase the storage capacity of whatever backing filesystem is |
| 30 |
hosting their files. |
| 31 |
|
| 32 |
> > All Joost is saying is that most resources can be overcommitted, since |
| 33 |
> > all the users will not be using all their resources at the same time. |
| 34 |
> |
| 35 |
> How do you overcommit disk space and then shrink the VMs automatically |
| 36 |
> when disk usage gets lower again? |
| 37 |
> |
| 38 |
|
| 39 |
Sorry, my previous example was bad, since the normal strategy is to |
| 40 |
expand when necessary as far as I know. See above. |
| 41 |
|
| 42 |
Alec |
Archives