Issue with quota

[Leicas]

I've seen severall mention online of issues with btrfs quotas, is there a way to make buttersink without using quota ?

I'm trying to sink to a remote ssh and by activating quota on that machine it almost always crash at some point with btrfs-cleaner using all cpu and memory... (lot of snapshots and arround 12TB of storage).

[eugene-bright]

It's known btrfs feature (cough, cough). See #38

You should try Linux kernels >=4.20

[eugene-bright]

There is also experimental branch (#60) of buttersink that uses timestamps to select base snapshots and do not rely on quotas. I do not have time to bring it to the master.

[Leicas]

Thanks, I saw the experimental branch but I wasn't sure if it works for ssh remote, and should I use the experimental branch on both the source and remote ?

As for switching to kernels >=4.20, sadly it's not available on debian stretch now and I'm not willing to go experimental on that machine...

Le lun. 18 mars 2019 à 05:18, eugene-bright [email protected] a écrit :

There is also experimental branch (#60 https://github.com/AmesCornish/buttersink/pull/60) of buttersink that uses timestamps to select base snapshots and do not rely on quotas. I do not have time to bring it to the master.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/AmesCornish/buttersink/issues/68#issuecomment-473829697, or mute the thread https://github.com/notifications/unsubscribe-auth/AAuwQzVVSgnAz7jfqBR6eg4LG5udkvCFks5vX1nzgaJpZM4b4kb3 .

[eugene-bright]

If you are going to give my branch a try you should update both sides as new field was introduced in RPC return record (otime). But you should use it on your own risk. A lot of things is known to be broken.