Ignored is the wrong word, it is not even relevant to the COMPONENT_RESTART use
case. The saAmfSUFailover configuration applies to COMPONENT_FAILOVER and
controls whether to fail or switch-over other components in the same SU.
---
** [tickets:#98] AMF: support for saAmfSUFailover**
**Status:** unassigned
**Created:** Mon May 13, 2013 05:07 AM UTC by Nagendra Kumar
**Last Updated:** Mon Jun 17, 2013 06:10 AM UTC
**Owner:** Praveen
Migrated from http://devel.opensaf.org/ticket/2701
As per AMF Spec: 3.11.1.3.2 Fail-Over Recovery Action:
If the service unit is configured to fail over as a single entity
(saAmfSUFailover set to SA_TRUE), all other components of the service unit are
abruptly termi-nated, and all service instances assigned to that service unit
are failed over;
But as per current implementation AMF gracefully removes the assignments for
the non faulted components first and then does failover. This needs to be
corrected as per spec
------------------------------------
Changed 11 months ago by hafe ¶
Duplicate of a classic?
http://devel.opensaf.org/ticket/6
Would be cooler to fix ticket number 6...
Changed 3 months ago by nagendra ¶
Some thoughts for the implementation:
0. It could be implemented for 2N to start with.
1. When saAmfSUFailover is set to TRUE and if component faults with below
recommended recovery, the behaviour would be as stated below:
Recommended Recovery Behaviour
--------------------------------------------------------------------------------
a. NO_RECOMMENDATION -> SU Failover
b. COMPONENT_RESTART -> SU Failover
c. COMPONENT_FAILOVER -> SU Failover
d. NODE_SWITCHOVER -> Do SU failover for the faulted SU and do
NODE SWITCHOVER to other SUs.
e. NODE_FAILOVER, NODE_FAST -> Ignore SU Failover
--------------------------------------------------------------------------------
2. If saAmfSUFailover is set to true, Consider Su failovers for escalation when
it is first reported or reported in between different levels of escalations. I
mean it could be treated as component faulted with recommended recovery as
SA_AMF_COMPONENT_FAILOVER for all escalation purposes.
3. Issues related to saAmfSUFailover set to true in existing imm.xml for
middleware 2N and No Red models need to be addressed in upgrade scenarios.
Modification of saAmfSUFailover need to be implemented and the modification
should be sent to amfnd from amfd. This will help in disabling saAmfSUFailover
flag after upgrade because going forward, if middleware cpnd/glnd crashes, we
need not do SU Failover.
Changed 7 weeks ago by hafe ¶
■summary changed from AVSv: su failover needs to be implemented as per the
spec to AMF: support for saAmfSUFailover
Changed 5 weeks ago by praveenmalviya ¶
■owner changed from ravisekhar to praveenmalviya
■status changed from new to accepted
■milestone changed from future_releases to 4.4.GA
I am accepting it only for 2N Model.
Changed 3 weeks ago by praveenmalviya ¶
I will consider NoRed? model along with 2N model.
---
Sent from sourceforge.net because you indicated interest in
<https://sourceforge.net/p/opensaf/tickets/98/>
To unsubscribe from further messages, please visit
<https://sourceforge.net/auth/subscriptions/>
------------------------------------------------------------------------------
This SF.net email is sponsored by Windows:
Build for Windows Store.
http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
Opensaf-tickets mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensaf-tickets
