Current events

From ClusterLabs
Jump to: navigation, search

Here is the latest news from the Pacemaker project

Fortnight ending December 28, 2007

Although not a lot of active work is being done at the moment, due to the time of year, I've been running tests as a background task and things are looking good for a release in mid-January. For more information about releases and release criteria, please visit the Releases and Release Testing pages.

I was initially concerned with the behavior of one cluster that likes to lock-up occasionally - however the problem seems rooted in VMware Server (virtual nodes will occasionally completely stop responding) rather than anything related to the cluster stack.

So far testing has been focused on the Heartbeat (2.1.3) stack but that should be completed in the few days at which point we'll begin testing on OpenAIS.

If anyone has a cluster they're not using and would like to help out with testing - please let me know.

Commit Highlights

  • High: Core: Fix peer-process tracking for the Heartbeat stack
  • High: Core: Heartbeat doesn't send status notifications for nodes that were already part of the cluster. Fake them instead
  • High: PE: Parse fail-count correctly (it may be set to ININFITY)
  • High: TE: Ensure failcount is set correctly for failed stops/starts
  • High: TE: Update failcount for oeprations that time out
  • High: crmd: Handle all possible peer states in crmd_ha_status_callback()
  • High: crmd: Make sure the current date/time is set - prevents use-of-NULL when evaluating rules
  • High: crmd: Supply the dispatch and destroy callback functions in the correct order
  • High: crmd: Use (node-processescrm_proc_ais) to accurately update the CIB after replace operations
  • Medium: AIS: Add a version field to the 'node-id' message
  • Medium: PE: Prevent a potential use-of-NULL in cron_range_satisfied()
  • Medium: crm_mon: Indicate when a node is both in standby mode and offline

--Abeekhof 20:40, 29 Dec 2007 (CET)

Week ending December 14, 2007

  • Most importantly, we selected a name: Pacemaker
  • Merged the AIS port
  • Purged almost all the heartbeat code from the source repository
  • Set up mailing lists (http://list.clusterlabs.org) and bugzilla (http://bugs.clusterlabs.org)
  • The choice of cluster stack is now selected automatically based on who started us
  • Many changes to allow supporting either or both cluster stacks
  • Ported attrd to run on OpenAIS
  • Re-evaluated heartbeat package layout
    • Stop building the built-in CRM
    • Split into essential (used by both stacks) and non-essential (only used by the heartbeat stack) components
  • New home for Pacemaker source code:

http://hg.clusterlabs.org/pacemaker/dev/

  • Temporary home for packages:

http://download.opensuse.org/repositories/home:/beekhof

  • Packages come in 3 flavors depending on which stacks are supported,
    • pacemaker supports both stacks
    • pacemaker-ais supports only OpenAIS
    • pacemaker-heartbeat supports only Heartbeat

Packages exist for current RedHat and SUSE distros. Debian and Ubuntu will follow next week.

This leaves us in pretty good shape for our first week. Starting Monday I will begin release testing for our first (feature frozen) stable series.

The plan is to have this completed by mid-January (not a lot will get done over the Christmas/New Year break) and provide bug-fix-only updates until at least 2009.

--Abeekhof 10:01, 16 Dec 2007 (CET)