Skip to content

Andrew Beekhof beekhof

Organizations

Cluster Labs matahari
April 15, 2014
beekhof commented on pull request ClusterLabs/pacemaker#501
Andrew Beekhof

Why is this not just: new_rsc_ref = create_xml_node(set, XML_TAG_RESOURCE_REF); crm_xml_add(new_rsc_ref, XML_ATTR_ID, obj_ref); New items always g…

April 15, 2014
Andrew Beekhof
  • Andrew Beekhof d4e1fb5
    Merge pull request #499 from gao-yan/expire-non-blocked-failure
  • Gao,Yan 30ea303
    Test: pengine: Ignore failure-timeout only if the failed operation ha…
  • 2 more commits »
April 15, 2014
beekhof merged pull request ClusterLabs/pacemaker#499
Andrew Beekhof
Fix: pengine: Ignore failure-timeout only if the failed operation has on-fail="block"
3 commits with 291 additions and 28 deletions
April 15, 2014
Andrew Beekhof
  • Andrew Beekhof 92d476c
    Merge pull request #475 from gao-yan/site-specific-params
  • Gao,Yan 3f298af
    Test: pengine: Support site-specific instance attributes based on rules
  • 1 more commit »
April 15, 2014
beekhof merged pull request ClusterLabs/pacemaker#475
Andrew Beekhof
Feature: pengine: Support site-specific instance attributes based on rules
2 commits with 149 additions and 0 deletions
April 15, 2014
beekhof commented on pull request ClusterLabs/pacemaker#475
Andrew Beekhof

And I do like simple :) Btw, I'm now on vacation until next wednesday.

April 15, 2014
Andrew Beekhof
April 15, 2014
Andrew Beekhof
April 15, 2014
Andrew Beekhof
  • Andrew Beekhof cfd845f
    Fix: PE: Do not initiate fencing for unclean nodes when fencing is di…
April 15, 2014
Andrew Beekhof
April 15, 2014
Andrew Beekhof
  • Andrew Beekhof 9f292bb
    Fix: crm_node: Only remove the named resource from the cib
  • Andrew Beekhof 6f441eb
    Test: PE: Unfencing is based on device probes, there is no need to un…
  • 15 more commits »
beekhof deleted branch rhel-7.0-rhpkg at beekhof/pacemaker
April 14, 2014
April 14, 2014
April 14, 2014
beekhof commented on pull request ClusterLabs/pacemaker#475
Andrew Beekhof

Would it not be better to store it in //crm_config (ie. where the software version and stack type are kept)?

April 14, 2014
beekhof commented on pull request ClusterLabs/pacemaker#499
Andrew Beekhof

Probably the best we can do

April 14, 2014
beekhof commented on pull request ClusterLabs/pacemaker#499
Andrew Beekhof

But also the only place that needs should_block right?

April 14, 2014
Andrew Beekhof
April 14, 2014
beekhof merged pull request ClusterLabs/pacemaker#463
Andrew Beekhof
Feature: pengine: Support cib object tags
2 commits with 2,149 additions and 201 deletions
April 14, 2014
beekhof commented on pull request ClusterLabs/pacemaker#463
Andrew Beekhof

I shudder to think :)

April 14, 2014
beekhof commented on pull request ClusterLabs/pacemaker#499
Andrew Beekhof

(/me looks more closely at the code) I was trying to avoid the secondary xpath lookup... Why not pass op to get_failcount_full() via check_operatio…

April 14, 2014
Andrew Beekhof
April 14, 2014
beekhof merged pull request ClusterLabs/pacemaker#495
Andrew Beekhof
Fix: crmd: Correctly detect changes of ticket state
1 commit with 2 additions and 2 deletions
April 14, 2014
beekhof commented on pull request ClusterLabs/pacemaker#463
Andrew Beekhof

One last question, why obj_ref here instead of resource_ref as we have for tickets?

April 14, 2014
beekhof commented on pull request ClusterLabs/pacemaker#499
Andrew Beekhof

I don't doubt this works but it feels like something we should be handling during unpack_rsc_op()

April 11, 2014
beekhof commented on pull request ClusterLabs/pacemaker#500
Andrew Beekhof

1.1 should have been merged back into master after the release. There shouldn't be any need to do cherrypicking.

April 11, 2014
Andrew Beekhof
April 11, 2014
beekhof merged pull request ClusterLabs/pacemaker#496
Andrew Beekhof
Test : cts : automatically determine whether to run RemoteBaremetal
1 commit with 10 additions and 0 deletions
April 11, 2014
beekhof commented on pull request ClusterLabs/pacemaker#496
Andrew Beekhof

Makes sense

April 11, 2014
Andrew Beekhof
  • Andrew Beekhof 6f441eb
    Test: PE: Unfencing is based on device probes, there is no need to un…
April 10, 2014
Andrew Beekhof
  • Andrew Beekhof 7aad7cb
    Fix: PE: Unfencing is based on device probes, there is no need to unf…
Something went wrong with that request. Please try again.