x/development-proposals
Code Issues Proposed changes

declare data center / region failures out of scope

Browse Source 
This would be way too ambitious for this user story.

Change-Id: Ia3f79659d47785fb2de6b51edfc896d3cc3f3210
Signed-off-by: Adam Spiers <aspiers@suse.com>
This commit is contained in:
Adam Spiers
2016-05-24 16:38:47 +01:00
parent 196c19c44f
commit d2d17f2f13
1 changed files with 13 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
user-stories/proposed/ha_vm.rst
View File

@@ -62,7 +62,19 @@ can be detected and recovered by the system. Possible failure events include:
* Attached Cinder volume failure
* Availability Zone/Data Center/Region failure
* Availability Zone failure
* Data Center / Region failure
Failure of a whole region or data center is obviously much more severe,
requiring recovery of not just compute nodes but also OpenStack services in
the control plane. It needs to be covered by a Disaster Recovery plan,
which will vary greatly for each cloud depending on its architecture,
supported workloads, required SLAs, and organizational structure. As such,
a general solution to Disaster Recovery is a problem of considerable
complexity, therefore it makes sense to keep it out of scope for this user
story, which should instead be viewed as a necessary and manageable step on
the long road to that solution.
N.B. This user story concerns high availability, not 100% availability.
Therefore some service interruption is usually expected when failures occur.