On Wed, Sep 23, 2015 at 03:28:29PM +0200, David Sterba wrote: > On Tue, Sep 22, 2015 at 03:39:30PM +0000, Hugo Mills wrote: > > > The way I would expect things to work is that a new subvolume > > > inherits it's properties from it's parent (if it's a snapshot), > > > > Definitely this. > > > > > or > > > from the next higher subvolume it's nested in. > > > > I don't think I like this. I'm not quite sure why, though, at the > > moment. > > I don't like inheritance from other than the parent subvolume because > this makes things less obvious. > > > It definitely makes the process at the start of allocating a new > > block group much more complex: you have to walk back up through an > > arbitrary depth of nested subvols to find the one that's actually got > > a replication policy record in it. (Because after this feature is > > brought in, there will be lots of filesystems without per-subvol > > replication policies in them, and we have to have some way of dealing > > with those as well). > > > > With an FS default policy, you only need check the current subvol, > > and then fall back to the FS default if that's not found. > > That looks reasonable to me. > > > These things are, I think, likely to be lightly used: I would be > > reasonably surprised to find more than two or possibly three storage > > policies in use on any given system with a sane sysadmin. > > Agreed. At the moment I'm thinking about all the configuration > possibilites we want to give to the users. Eg. the inheritance can be > configurable on the property level. > > The usecase: the toplevel has compression enabled but I don't want any > new subvolume share this property automatically. > > (The blockgroup type is probably a bad example for configurable > inheritance as it would not work for shared extents if the type is > different.)
There's some major questions about the semantics and the UI for per-subvol RAID, too: (Assume that A is RAID-1) # btrfs sub snap A A' # btrfs prop set A' raid raid-5 What happens? Do we convert A' to RAID-5 immediately, doubling (or more) the disk usage? Do we do lazy conversion, so that only new data in A' is written to RAID-5? Do we need a "btrfs sub restripe A'" command? Is directly setting properties actually the right interface to use here? (Reading properties... yes, but we still have the possible issues of getting information on partially-converted subvols). # btrfs sub snap A A' # A' is also RAID-1 here: inherit from A # btrfs prop set A' raid raid-5 # btrfs sub restripe A' (you wait, time passes) # btrfs sub snap A A'' # btrfs prop set A'' raid raid-5 # btrfs sub restripe A'' How do I make A' and A'' pick up their sharing again? Dedup? Will this cause problems with incremental send? # btrfs balance start -dconvert=single A What does this do? Convert the whole FS's data to single? Just subvol A? If the whole FS, will all the subvols' raid xattrs be dropped? Or will only the FS-default ones be changed? If only the FS-default xattrs are changed, then a write to any non-default subvol will revert to its configured RAID level, and we almost immediately lose the "everything single" configuration requested by the convert... # btrfs balance start -mconvert=single A What does this do? (Probably convert the whole FS's metadata to single -- which may well answer the previous question as well, unless we want to surprise users). I'll probably think of some more questions later, but that's enough for going on with. :) Hugo. -- Hugo Mills | 2 + 2 = 5, for sufficiently large values of 2. hugo@... carfax.org.uk | http://carfax.org.uk/ | PGP: E2AB1DE4 |
signature.asc
Description: Digital signature
