Steps to reproduce:
1) Bring 2N application with all NPI SUs with each SU having 2 components. AMF performs
passive monitoring of components.
2) Perform admin lock on active SU.
3) When termination of first component is in progress, comp2 should fault and cleaned up
successfully. Here AMFND moves SU to UNINSTANTIATED state even though comp1 has not
responded for terminate command.
4) Now when comp1 responds for terminate command, AMFND does not take any action as SU is
already in unassigned state. Here AMFND should respond to AMFD for assignment.
5)Since AMFD does not get assignment response from AMFND, it does not perform fail-over.
Will upload traces and configuration.
Traces
changeset: 5091:645a5b35c100
branch: opensaf-4.3.x
parent: 5082:f96a9b16283c
user: praveen.malviya@oracle.com
date: Fri Mar 28 10:43:13 2014 +0530
summary: amfnd: fix failover of NPI SU during SU lock with fault [#820]
changeset: 5092:90d97fea11dd
branch: opensaf-4.4.x
parent: 5085:560e4191aa5d
user: praveen.malviya@oracle.com
date: Fri Mar 28 10:44:34 2014 +0530
summary: amfnd: fix failover of NPI SU during SU lock with fault [#820]
changeset: 5093:fa608e5fe060
tag: tip
parent: 5090:cbd634d69f62
user: praveen.malviya@oracle.com
date: Fri Mar 28 10:45:03 2014 +0530
summary: amfnd: fix failover of NPI SU during SU lock with fault [#820]
[staging:645a5b]
[staging:90d97f]
[staging:fa608e]
Related
Tickets:
#820