Xfactor973 Gluster Charm
- By Chris Holcombe
- Storage
Channel | Revision | Published | Runs on |
---|---|---|---|
latest/stable | 6 | 19 Mar 2021 | |
latest/edge | 1 | 19 Mar 2021 |
juju deploy xfactor973-gluster-charm
Deploy universal operators easily with Juju, the Universal Operator Lifecycle Manager.
Platform:
-
cifs | boolean
Enable CIFS exporting of the volume. This option will enable windows clients to access the volume.
-
cluster_type | string
Default: Replicate
The type of volume to setup. Replicate is sufficient for most use cases. Other options include: Distribute, Stripe, StripedAndReplicate, Disperse, DistributedAndStripe, DistributedAndReplicate, DistributedAndStripedAndReplicate, DistributedAndDisperse.
-
filesystem_type | string
Default: xfs
The filesystem type to use for each one of the bricks. Can be either zfs, xfs, btrfs, or ext4. Note that zfs only works with ubuntu 16.04 or newer
-
key | string
Key ID to import to the apt keyring to support use with arbitary source configuration from outside of Launchpad archives or PPA's.
-
replication_level | int
Default: 3
This sets how many replicas of the data should be stored in the cluster. Generally 2 or 3 will be fine for almost all use cases. Greater than 3 could be useful for read heavy uses cases.
-
source | string
Default: ppa:gluster/glusterfs-3.8
Optional configuration to support use of additional sources such as: - ppa:myteam/ppa - cloud:trusty-proposed/kilo - http://my.archive.com/ubuntu main The last option should be used in conjunction with the key configuration option. NOTE: Changing this configuration value after your cluster is deployed will initiate a rolling upgrade of the servers one by one.
-
splitbrain_policy | string
Default: size
Split brain means that the cluster can not come to consensus on which version of a file to serve to a client. This option set automatic resolution to split-brains in replica volumes Options include: ctime|mtime|size|majority. Set this to none to disable. Example: Setting this to "size" will pick the largest size automatically and delete the smaller size file. "majority" picks a file with identical mtime and size in more than half the number of bricks in the replica.
-
sysctl | string
Default: { vm.vfs_cache_pressure: 100, vm.swappiness: 1 }
YAML-formatted associative array of sysctl key/value pairs to be set persistently. By default we set pid_max, max_map_count and threads-max to a high value to avoid problems with large numbers (>20) of OSDs recovering. very large clusters should set those values even higher (e.g. max for kernel.pid_max is 4194303).
-
virtual_ip_addresses | string
By default NFSv3 is installed and started on all servers. If the server the client is connected to dies NFSv3 doesn't have a failover mechanism. Setting this option to a space separated list of ip addresses cidr's to be used as virtual ip addresses on the cluster such as: - 10.0.0.6/24 10.0.0.7/24 10.0.0.8/24 10.0.0.9/24 With a cluster of 2 machines would assign 2 virutal addresses to each host. - 2001:0db8:85a3:0000:0000:8a2e:0370:7334/24 2001:cdba:0000:0000:0000:0000:3257:9652/24 With a cluster of 2 machines would assign 1 virtual address to each host. Each server will be assigned (# of ip addresses) / ( # of servers) ip addresses to use for quick failover. The administrator is expected to create a DNS A record with the virtual ip addresses before setting this. When clients connect to this charm it will attempt to resolve its own virtual ip and hand out the DNS address instead of the ip address if possible. Upon failure of a server ctdb will issue a gratitous arp requests and migrate the virtual ip address over to a machine that is still up. Note however that there are implications here for load balancing and moving an additional ip address to a server could increase load on it.
-
volume_name | string
Default: test
The name of the Gluster volume to create. This will also serve as the name of the mount point. Example: mount -t glusterfs server1:/test