Here are updated versions of my direct data sources I put together (communities doesn't seem to let me edit the existing article) - the 20150703 update adds support for WFA 3.0 and provides much more complete cDOT (cm_storage), including cDOT 8.3 features. The 20150728 update fixes storage pool support. The 20150802 update adds support for WFA 3.1RC1 and fixes failover_group support in WFA 3.0.
Not everyone has pre-existing OCUM and installing and configuring OCUM for both 7mode and cDOT is not necessarily a small task. These datasource types populate the standard storage and cm_storage schemas which can then be used as normal with many workflows.
The WFA 3.0 cDOT direct data source is almost complete. It now populates the following: disk, disk_aggregate, efficiency_policy. It includes cDOT 8.3 information added in WFA 3.0: broadcast_domain, failover_group, ipspace, storage_pool*. There may be some minor gaps in some of the data - please let me know if you identify anything. (Note: The WFA 2.2* cDOT data source is missing all of these items).
The WFA 3.1RC1 cDOT dats source adds support for cluster_peer, vserver_peer and SVM DR, and all new fields added in 3.1RC1 are populated. It also populates cifs domain information.
The 7-mode source is less complete. It contains: array, vfiler, aggregate, interface, volume, vsm, qtree, lun. It is missing: array_license, cifs_share, cifs_share_acl, dataset, disk, igroup, igroup_initiator, lunmap, nfs_export, object_comment, quota, snapshot, snapvault, user_quota.
Configuration is simple:
- Provide credentials for the arrays/clusters in WFA credentials.
- Add datasource - select version appropriate to your WFA version (the 7-mode source works with 2.2-3.0) which the cDOT one is different for 2.2 and 3.0.
- Set the Hostname of the datasource to a comma separated list of array or cluster admin names or addresses, or set up one datasource per array/cluster.
- Configure the interval to something reasonable (say 15 or 30 minutes)
- Increase the timeout for larger environments (in my testing, collection takes 15-90s per array/cluster - it may take longer when there are large numbers of objects or if there is significant latency between the WFA system and the array/cluster)
There are some flags that can be passed in (comma separated) through the data source's "User name" field:
- strict - fail collection if any error occurs (default if there is only one array/cluster)
- nostrict - don't fail connection if an error occurs (for a single array/cluster datasource)
- debug - log more information
- log=<logpath> - where to log (defaults to ..\..\log\direct_<schema>.log - typically C:\Program Files\NetApp\WFA\jboss\standalone\log\)
- timeout=<timeout_in_milliseconds> - API timeout (defaults to 180000)
The Password, Database and Port fields are unused.
Reservations, etc, should work as normal. This has been developed against WFA 2.2RC1, 2.2, 2.2.1, 3.0RC1 and 3.0 and will likely require updates to work with future versions.
Upgrades from previous versions should be seamless.
Changes from previous version:
- Added WFA 3.0 schema support
- Accomodate changed mysql behavior in WFA 3.0
- Single array/cluster is now automatically strict
- Added nostrict parameter
- 7-mode direct data source handles pre 7.3.3 systems
- 7-mode direct data source doesn't fail if vfiler stopped/inconsistent
- Fix for storage_pools in 3.0 cDOT data source (20150728)
- Fix for broadcast_domain (3.0/3.1RC1)
- Added support for WFA 3.1RC1, populates all new schema tables and fields.
- No longer imports into WFA 2.2RC1
