Menu

#64 RDE plugin support for arbitration at startup

5.18.02
fixed
nobody
None
enhancement
rde
-
major
False
2018-08-24
2013-05-08
No

Migrated from Ticket #2478

Two controllers starting up without being able to communicate with each other could get into a split brain situation - both are active. If these controllers share a disk, data corruption could be the result.

RDE should support some plugin mechanism to allow for arbitration

Note: This feature has been evolved in subsequent tickets; see tickets [#2795] and [#2918]

1 Attachments

Related

Tickets: #2795
Tickets: #2918
Wiki: ChangeLog-5.18.02
Wiki: NEWS-5.18.02

Discussion

  • Anders Widell

    Anders Widell - 2017-12-14
    • summary: RDE plugin support for resource fencing at startup --> RDE plugin support for arbitration at startup
    • Description has changed:

    Diff:

    --- old
    +++ new
    @@ -1,5 +1,5 @@
     Migrated from Ticket #2478
    
    -Two controllers without communication could get into a split brain situation - both are active. If these controllers share a disk, data corruption could be the result.
    +Two controllers starting up without being able to communicate with each other could get into a split brain situation - both are active. If these controllers share a disk, data corruption could be the result.
    
    -RDE should support some plugin mechanism to allow for fencing a node from a shared resource
    +RDE should support some plugin mechanism to allow for arbitration
    
    • Attachments has changed:

    Diff:

    --- old
    +++ new
    @@ -0,0 +1 @@
    +Split-brain prevention in OpenSAF.pdf (87.5 kB; application/pdf)
    
    • Blocker: --> False
    • Milestone: future --> 5.18.01
     
  • Gary Lee

    Gary Lee - 2018-01-26

    Updated Overview PR. Section 3.7.6 added.

     
  • Gary Lee

    Gary Lee - 2018-01-30

    commit 23c6d1638fdf73db1dff5127014a06a049cac814
    Author: Gary Lee gary.lee@dektech.com.au
    Date: Thu Jan 25 14:44:36 2018 +1100

    doc: update README [#64]
    

    commit 685ab1cc41a12015dfaf008720b421c6c48ec26d
    Author: Gary Lee gary.lee@dektech.com.au
    Date: Thu Jan 25 14:44:36 2018 +1100

    fmd: update consensus service during controller failover [#64]
    

    commit 22a1bc7e9b5d22fea41876caf52a6e2104c4c956
    Author: Gary Lee gary.lee@dektech.com.au
    Date: Thu Jan 25 14:44:36 2018 +1100

    amfd: update consensus service when performing SI swap [#64]
    
    When a node goes down and split-brain prevention is enabled,
    check that we still have write access to the consensus service.
    If not and fencing is disabled, reboot the node to prevent
    split brain.
    

    commit cd9a2c5754047908f3ccb7bfec2e144ff0c4618e
    Author: Gary Lee gary.lee@dektech.com.au
    Date: Thu Jan 25 14:44:36 2018 +1100

    rded: add split brain prevention support [#64]
    
    * consult with consensus service before promoting node to active
    * add watch thread and self-fence if it detects active controller
      has been changed (if remote fencing is disabled)
    

    commit 522b18114ed37184479f1012ed9fde6a7feefa6d
    Author: Gary Lee gary.lee@dektech.com.au
    Date: Thu Jan 25 14:44:36 2018 +1100

    osaf: add consensus API [#64]
    
     
  • Gary Lee

    Gary Lee - 2018-01-30
    • status: unassigned --> fixed
     
  • Anders Widell

    Anders Widell - 2018-08-24
    • Description has changed:

    Diff:

    --- old
    +++ new
    @@ -3,3 +3,5 @@
     Two controllers starting up without being able to communicate with each other could get into a split brain situation - both are active. If these controllers share a disk, data corruption could be the result.
    
     RDE should support some plugin mechanism to allow for arbitration
    +
    +Note: This feature has been evolved in subsequent tickets; see tickets [#2795] and [#2918] 
    
     

    Related

    Tickets: #2795
    Tickets: #2918


Log in to post a comment.