- **status**: review --> fixed
- **Comment**:
default(4.6):
changeset: 6174:ceefbeb5c1c4
user:Hung Nguyen
date:Mon Nov 17 15:00:48 2014 +0700
summary: immtools: Add wrapper functions for saImmOmCcbValidate and
saImmOmCcbAbort [#1200]
changeset: 6175:69075a7df35b
tag:
---
** [tickets:#1223] pyosaf: add immom package (from opensaf-tools)**
**Status:** unassigned
**Milestone:** 4.6.FC
**Created:** Fri Nov 28, 2014 08:32 AM UTC by Hans Feldt
**Last Updated:** Fri Nov 28, 2014 08:32 AM UTC
**Owner:** nobody
It is proposed to move the package found in repo open
- **status**: unassigned --> accepted
- **assigned_to**: Anders Bjornerstedt --> Neelakanta Reddy
- **Part**: - --> nd
---
** [tickets:#1091] 2PBE: class create timesout before default SYNCR_TIMEOUT**
**Status:** accepted
**Milestone:** 4.6.FC
**Created:** Mon Sep 15, 2014 03:45 PM UTC by Siri
- **Comment**:
Re-opening the ticket to fix a defective NULL initialize in the original fix
---
** [tickets:#1221] clm: protect node down processing when checkpointing arrives
earlier**
**Status:** fixed
**Milestone:** 5.0
**Labels:** faster checkpointing failover [#1120]
**Created:** Thu N
- **Comment**:
The original fix had a couple of reviews. the initial patch breaked out of the
while loop, but the final pushed version removed the break but missed out in
removing the defective NULL initialize.
The following changesets fixes this problem. Now, it works as expected.
changeset:
- **Comment**:
Part 3 and final fixes. Anything else would be a separate ticket.
Tested with 10 failovers and 10 switchovers intermixed with payload reboots.
changeset: 6179:79a81a332393
branch: opensaf-4.4.x
parent: 6176:cd31a75a0059
user:Mathivanan N.P.
date:Fri Nov
- Description has changed:
Diff:
--- old
+++ new
@@ -10,6 +10,12 @@
It is important that *everyone* understands this.
There is no magic being done here.
-Use case: opensaf cloud deployment
+Use case: opensaf cloud deployment. In a cloud deployment, the risk for
multiple simultaneous no
I think we cannot simply expand mbcsv to multiple standbys as it is. The
problem lies in how to deal with failures: what to do if one of the standby
nodes is very slow or doesn't reply at all. How to handle the case when the
active fails at a point when only a subset of the standbys have receive