Feature #1790
closed
rbd: have a way of establishing configured mappings at boot time
Added by Anonymous over 12 years ago.
Updated over 10 years ago.
Description
We need to be careful about the config format, to make automatic editing easy (think Chef).
First draft:
/etc/ceph/rbd.d/NAME.conf is used as ceph.conf to map image called NAME.
You can drop in multiple .conf files in that directory.
Two major use cases:
1)
ln -s /etc/ceph/ceph.conf /etc/ceph/rbd.d/foo.conf
maps image "foo" at boot time with the main ceph.conf settings, with the keyrings configured there
2)
install -m0600 appropriate-keyring /etc/ceph/rbd.d/bar.keyring
cat >>/etc/ceph/rbd.d/bar.conf <<EOF
[global]
...
keyring = bar.keyring
EOF
What if your image is not in the pool "rbd" ?
I was thinking about a 'rbdtab' file:
#Monitor(s) #Args #Pool #Image
192.168.1.6 id=admin,secret=/etc/ceph/secret rbd alpha
192.168.1.6 id=admin,secret=/etc/ceph/secret rbd2 beta
Wouldn't that be easier?
Single-file configuration is more annoying to handle with automated tools, file-per-device gives you good atomicity on manipulating the configuration.
Your point about pools in very good. I originally thought of putting "[rbd foo]" etc inside the config file, but I'm really hesitant to add special-purpose, more dynamic, fields/sections into the config file, and feel this is on a different level from ceph.conf anyway.
Perhaps /etc/ceph/rbd.d/ANYTHINGUNIQUEHERE.conf with contents like
image = NAME # maybe leave out to take from name of conf file
pool = notrbd # leave out for default "rbd"
id = cephkeythatcanmaprbd # leave out for default "admin"
config = /etc/ceph/non-default-ceph.conf # leave out for default
- Project changed from Ceph to rbd
- Status changed from New to Resolved
commit:a4ddf704868832e119d7949e96fe35ab1920f06a
Also available in: Atom
PDF