Installs an NFS server that other charms can relate to in order
to create exports and attach to the NFS storage
This is the NFS charm with support for trusty, xenial, and bionic. It creates an NFSv4(http://nfs.sourceforge.net/)
server using the local filesystem as a backing store. There is no clustering or
federation, so multiple units would result in multiple nfs servers. This is NOT
a highly available server.
To deploy an NFS server:
juju deploy nfs
Each relation to the NFS charm will result in a new export for the units in that
application. The default location of data is /srv/data
Note that the NFS charm will NOT destroy data. This means if you remove a relation
to an application, you will need to manually destroy the data saved by that application
in the local data location.
conjure-up canonical-kubernetes juju deploy nfs juju add-relation nfs kubernetes-worker
juju deploy nfs juju deploy mysql juju deploy owncloud juju add-relation mysql:owncloud juju add-relation nfs:nfs owncloud:shared-fs
The above example deploys OwnCloud personal cloud storage, and provides remote storage via the NFS host.
juju deploy nfs juju deploy mysql juju deploy wordpress juju add-relation mysql:db wordpress:db juju add-relation nfs:nfs wordpress:nfs
To migrate storage from one NFS unit to another, first add the new unit in
such a way as to avoid publishing it before it's ready:
juju config nfs active_units=<old unit ID> juju add-unit nfs
Now start the downtime:
juju config nfs active_units=none
Wait for all clients to unmount, then move the underlying storage to the new
unit in whatever way is appropriate for your deployment.
Finish the downtime by publishing the new unit:
juju config nfs active_units=<new unit ID>
After clients have mounted the new unit and you've checked that all is well,
you can remove the old unit:
juju remove-unit <old unit ID> juju config nfs active_units=
Note that the migration process is quite abrupt: the server does not wait
for all clients on related units to unmount before stopping. Consider
arranging separately for downtime of clients that might write to their NFS
Known Limitations and Issues
No high availability story
At present the charms consuming an NFS relationship only account for a single host. Most charms assume the first incoming NFS mount-point is the sole replacement, and subsequent NFS relationship-join requests are ignored.
If you are attempting to deploy NFS to an LXC container, such as the juju local provider, there are additional steps that need to be taken prior to deploying the NFS charm.
On the LXC host:
apt-get install nfs-common modprobe nfsd mount -t nfsd nfsd /proc/fs/nfsd
Edit /etc/apparmor.d/lxc/lxc-default and add the following three lines to it:
mount fstype=nfs, mount fstype=nfs4, mount fstype=nfsd, mount fstype=rpc_pipefs,
sudo /etc/init.d/apparmor restart
juju deploy nfs
- storage_root: The root path where exported directories will be created
- export_options: The default export options. Ships with rw,sync,no_root_squash,no_all_squash
- mount_options: The default client mount options
Mike Wilson email@example.com
Upstream NFS Project
- To view the source: https://github.com/hyperbolic2346/nfs-charm
- (string) If set, a comma-separated list of unit names that should publish data on the 'mount' relation. This makes it possible to transition gracefully between instances of this application.
- (string) The default export options
- (int) The number of nfs daemons to run on startup
- (string) The default client mount options
- (string) Used by the nrpe subordinate charms. A string that will be prepended to instance name to set the host name in nagios. So for instance the hostname would be something like: juju-myservice-0 If you're running multiple environments with the same services in them this allows you to differentiate between them.
- (string) A comma-separated list of nagios servicegroups. If left empty, the nagios_context will be used as the servicegroup.
- (string) The root path where exported directories will be created