| 1 |
On 14/08/18 12:42, John Covici wrote: |
| 2 |
> I use sanoid/syncoid to back up using zfs. Its great, keeps snapshots |
| 3 |
> for as long as I want them (I use 80 days for now). And it keeps |
| 4 |
> hourlies for the last couple of days as well, so I could roll back in |
| 5 |
> case of a problem. Very nice if you use zfs. I don't think btrfs is |
| 6 |
> ready for prime time -- its been under development for a while, but I |
| 7 |
> am scared to use it -- I did try once, but got nowhere. |
| 8 |
|
| 9 |
btrfs is pretty reliable apparently unless you stress it in ways where |
| 10 |
it is known to be flaky ... and it's now SUSE's default root filesystem ... |
| 11 |
|
| 12 |
Basically, DO NOT use raid, unless it's raid 1. DO NOT let the disk fill |
| 13 |
up if you're using snapshots. |
| 14 |
|
| 15 |
I think one of the reasons it's still got a bad rap, is a lot of SUSE |
| 16 |
installs fell over pretty badly. And the reason they fell over is that |
| 17 |
the developers forgot a lot of users don't have powerful machines ... |
| 18 |
|
| 19 |
The SUSE defaults were to enable snapshotting for upgrades and updates. |
| 20 |
And not to check the size of disk allocated to / ... :-( so it was not |
| 21 |
unusual for an update to snapshot, install, run out of disk and BOOOOM :-( |
| 22 |
|
| 23 |
Made worse by the fact that when they realised this, they didn't try |
| 24 |
to fix it retrospectively, leaving a lot of time-bombs out there just |
| 25 |
waiting to go off. |
| 26 |
|
| 27 |
Fortunately, when I hit this, I had spare disk I could add to the volume |
| 28 |
and it promptly sorted itself out, but a simple dedicated root should |
| 29 |
not fill up a 10GB partition (or was it 80? I'm not sure ...). But no, a |
| 30 |
lot of people have trashed their disks trying to recover from a disk |
| 31 |
full scenario :-( |
| 32 |
|
| 33 |
Cheers, |
| 34 |
Wol |
Archives