Ensure all nodes are re-unfenced after device configuration change
crm_resource --why now checks node health status
Resolves: rhbz1872483
Resolves: rhbz2065818

Add support for ACL groups
Resolves: rhbz1724310

Restore crm_attribute query behavior when attribute does not exist
Resolves: rhbz2072107

Fencer should not ignore CIB updates when stonith is disabled
Rebase pacemaker on upstream 2.1.4 final release
Fix typo in ocf:pacemaker:HealthSMART meta-data
Resolves: rhbz2055935
Resolves: rhbz2072107
Resolves: rhbz2094855

crm_attribute works on remote node command line when hostname differs
Rebase pacemaker on upstream 2.1.3 final release
Resolves: rhbz1384172
Resolves: rhbz2072107

crm_resource --restart fails to restart clone instances except instance 0
Add new multiple-active option for "stop unexpected instances"
Unable to show metadata for "service" agents with "@" and "." in the name
Resource ocf:pacemaker:attribute does not comply with the OCF 1.1 standard
Allow resource meta-attribute to exempt resource from node health restrictions
Show node health states in crm_mon
Rebase pacemaker on upstream 2.1.3-rc2 release
crm_mon API result does not validate against schema if fence event has exit-reason
Resolves: rhbz1930578
Resolves: rhbz2036815
Resolves: rhbz2045096
Resolves: rhbz2049722
Resolves: rhbz2059638
Resolves: rhbz2065812
Resolves: rhbz2072107
Resolves: rhbz2086230

Fix regression in down event detection that affects remote nodes
Resolves: rhbz2046446

Improve display of failed actions
Handle certain probe failures as stopped instead of failed
Update pcmk_delay_base description in option meta-data
Avoid crash when using clone notifications
Retry Corosync shutdown tracking if first attempt fails
Resolves: rhbz1470834
Resolves: rhbz1506372
Resolves: rhbz2027370
Resolves: rhbz2039675
Resolves: rhbz2042550

Correctly get metadata for systemd agent names that end in '@'
Use correct OCF 1.1 syntax in ocf:pacemaker:Stateful meta-data
Fix regression in displayed times in crm_mon's fence history
Resolves: rhbz2003151
Resolves: rhbz2027370
Resolves: rhbz2032027

Allow per-host fence delays for a single fence device
Use OCF 1.1 enum type in cluster option metadata for better validation
Allow spaces in pcmk_host_map
ACL group names are no longer restricted to a unique XML id
Rebase on upstream 2.1.2
Ensure upgrades get compatible Corosync libraries
Resolves: rhbz1082146
Resolves: rhbz1281463
Resolves: rhbz1346014
Resolves: rhbz1376538
Resolves: rhbz1384420
Resolves: rhbz2011973
Resolves: rhbz2027006

Fix XML issue in fence_watchdog meta-data
Resolves: rhbz1443666

Fix minor issue with crm_resource error message change
Resolves: rhbz1447918

Fix watchdog agent version information
Ensure transient attributes are cleared when multiple nodes are lost
Resolves: rhbz1443666
Resolves: rhbz1986998

Allow configuring specific nodes to use watchdog-only sbd for fencing
Resolves: rhbz1443666

Show better error messages in crm_resource with invalid resource types
Avoid selecting wrong device when dynamic-list fencing is used with host map
Do not schedule probes of unmanaged resources on pending nodes
Fix regressions in crm_attribute and crm_master argument handling
Resolves: rhbz1447918
Resolves: rhbz1978010
Resolves: rhbz1982453
Resolves: rhbz1984120

crm_resource now supports XML output from resource agent actions
Correct output for crm_simulate --show-failcounts
Avoid remote node unfencing loop
Resolves: rhbz1644628
Resolves: rhbz1686426
Resolves: rhbz1961857