Pacemaker 1.1.23 - Release Candidate 1

• Features added since Pacemaker-1.1.22

  • fencing: improve parameter handling for RHCS-style fence agents
    - add port or plug parameter according to metadata
    - plug parameter shown as non-required in the metadata
  • tools: add stonith_admin --delay option (default = 0) to support enforced fencing delay

• Changes since Pacemaker-1.1.22

  • crmd: properly detect node info requests from remote nodes (regression introduced in 1.1.19)
  • crmd: remove node as voter only if lost (regression introduced in 1.1.20)
  • crmd: clear leaving node's transient attributes even if there is no DC
  • crmd: don't ack same request twice
  • crmd: avoid delay when recovering resources after shutdown-lock expiration (BZ#5424)
  • fencing: remove dangling 'pending' for completed DC fencing (BZ#5401)
  • fencing: avoid possible use-of-NULL when parsing metadata
  • fencing: "action" now shown as not required in agent metadata
  • libcrmcommon: fix parsing ISO 8601 offsets with a plus sign
  • libcrmcluster: use proper type for corosync id when updating node state
  • pengine: don't incorporate dependencies' colocations multiple times
  • pengine: ensure attenuated scores still have some effect
  • pengine: ignore colocation constraints with 0 score
  • pengine: make sure cluster-wide maintenance-mode=true overrides per-resource settings

The priority-fencing-delay feature from the Pacemaker 2.0.4 release can be enabled for this release by building with -DENABLE_PRIORITY_FENCING_DELAY in CFLAGS, but be aware that unexpected behavior may result in mixed-version clusters or after an upgrade to a 2.0 version earlier than 2.0.4.