ZFS Jails: Recursively mounting ZFS dataset children?

N

NSLee0

Running 11.0-RELEASE, ZFS on root.

I created a jail, into which I put a few services/scripts to backup my filesystem to another local server, as well as the cloud.

I initially planned on using nullfs to mount my important datasets into this jail (read only), and have it read/backup the data from there. It seems however that each ZFS dataset needs to be individually mounted in the jail's fstab. I haven't been able to just mount /, and have all child datasets below it also be mounted.

E.g. if I have datasets for:
/
/jails
/jails/jail1
/jails/jail2
/home
/home/user1
/home/user2
...

I have to add each of those individually into my backup jail's fstab, I can't just add "/" and have "/jails/jail1", "/jails/jail2", "/home/user1" ... also mount.

I don't mind adding them all by hand once, but this seems like a nightmare for maintenance as each time I create a new dataset I also have to add it to the backup jail's mounts or else it won't get archived.

Is there a way to recursively mount all child datasets when mounting the parent?
 
host: 15.0-RELEASE

NSLee0 - thanks for posting. Here, 10 years later, and I'm doing the same thing you did.

my zfs mounts

Code: 
# zfs mount | grep opt
zroot/SAFE/dev                  /opt/dev
zroot/SAFE/stage                /opt/stage
zroot/SAFE/prod                 /opt/prod
zroot/SAFE/dev/stuff1            /opt/dev/stuff1
zroot/SAFE/stage/stuff1          /opt/stage/stuff1
zroot/SAFE/prod/stuff1           /opt/prod/stuff1
zroot/SAFE/stage/stuff2       /opt/stage/stuff2
zroot/SAFE/dev/stuff2         /opt/dev/stuff2
zroot/SAFE/prod/stuff2        /opt/prod/stuff2

my jail's fstab entry

Code: 
# /jails/fstab/foo.fstab
/opt/dev/stuff1    /jails/containers/foo/opt/dev/stuff1          nullfs ro 0 0
/opt/dev/stuff2    /jails/containers/foo/opt/dev/stuff2          nullfs ro 0 0

/opt/stage/stuff1    /jails/containers/foo/opt/stage/stuff1          nullfs ro 0 0
/opt/stage/stuff2    /jails/containers/foo/opt/stage/stuff2          nullfs ro 0 0

/opt/prod/stuff1    /jails/containers/foo/opt/prod/stuff1          nullfs ro 0 0
/opt/prod/stuff2    /jails/containers/foo/opt/prod/stuff2          nullfs ro 0 0

on the host, I touch the fakedata.txt and then view it in the jail

Code: 
# jexec -l foo find /opt/
/opt/
/opt/stage
/opt/stage/stuff1
/opt/stage/stuff1/fakedata.txt
/opt/stage/stuff2
/opt/prod
/opt/prod/stuff1
/opt/prod/stuff2
/opt/dev
/opt/dev/stuff1
/opt/dev/stuff2

If there's a more elegant way to mount all of this, feel free to chime in. For me, my setup is fairly static... but since I use Ansible to push changes to my fleet, changes aren't a big deal.
 
nullfs mounts are per-dataset and don't recurse (if they did, you could create an infinite loop in the filesystem tree by crossmounting nullfs mountpoints on each other); there's not a more elegant way to do this, unfortunately.
 
atax1a Thank you. Since you know about this stuff, can you look at a the bottom of this thread where I'm asking about how to mount stuff into a jail so regular (non root) users can see them?

toddg

jails Post in thread 'infrastructure as code and orchestration tools : ansible vs iocage vs (appjail + director + overlord)'

@victort bsd-apps looks awesome!!! I'll definitely be looking closely at that, as you have scripts for at least 4 things I need in there.

Here's my question on this: the README mentions: "Mount your directories and set your properties and variables as needed"... but how do you mount datasets such that the jail users have rw permissions to them?

For example, PostgreSQL is going to have a postgres user account that's going to need to be able to write the database files. And those files need to be written to a zfs dataset that was created on the HOST.

For context...
 
You must log in or register to reply here.
Back
Top