Re: [users] troubles getting HA Assignment

[Mohan K. <mohan@GetHighAvailability.com>]

Hi Jim,
We downloaded opensaf 5.2.0 and started SC-1(Active) and SC-2(Standby)
running 2N redundancy model Amf demo on both the controllers. Amf demo is
Act on SC-1 and Standby on SC-2. We performed the following commands to do
admin operations:
1. amf-adm lock   safAmfNodeGroup=AllNodes,safAmfCluster=myAmfCluster
[This node group contains SC-1 and SC-2 in imm.xml (default)]
   The assignments got removed from Amf demo instances.
2. amf-adm lock-in   safAmfNodeGroup=AllNodes,safAmfCluster=myAmfCluster

   Both demo instances got terminated.
3. amf-adm unlock-in   safAmfNodeGroup=AllNodes,safAmfCluster=myAmfCluster

   Both demo instances got started.
4. amf-adm unlock   safAmfNodeGroup=AllNodes,safAmfCluster=myAmfCluster    
   Amf demo instances got Act and Standby assignments.
We did immlist on SUs of Amf demo, they are all in IN-SERVICE with
assignments.
We also did some more configuration changes like: 
- Standby instance was hosted on payload(PL-3) and then performed the same
steps. Everything was as per expectation.
- Standby instance was hosted on payload(PL-3) but payload was down while
performing the above steps. Everything was as per expectation here as SU2
was never up and was always showing out of service.

So, we couldn't reproduce the issue reported. If we haven't performed the
steps, you did, then please correct our steps and we would try again. If
possible, can you please share the traces of Amfnd and Amfd and immdump
output when issue is reproduced. You can enable amf traces as below:
kill -USR2 <amfd pid>
kill -USR2 <amfnd pid>
And then run the steps to reproduce. You can disable the traces again by
running the same commands as mentioned above.

Also, you can check:
-	operational, administrative, and presence states of a service unit
-	operational state of its containing node
-	administrative states of its containing node, service group,
application, and the cluster
-	administrative state of ClmCluster

As per Amf Specs:
"The operational, administrative, and presence states of a service unit, the
operational
state of its containing node, and the administrative states of its
containing node, service
group, application, and the cluster are combined into another state, called
the
readiness state of a service unit."

Hope it helps!

Thanks & Regards
Mohan Kanakam | 91-8333082448
Senior Software Engineer
High Availability Solutions
 www.GetHighAvailability.com
Get High Availability Today !
NJ, USA: 1 508-507-6507    |    Hyderabad, India: 91 798-992-5293

-----Original Message-----
From: Carroll, James R [mailto:jam...@lm...] 
Sent: 26 January 2022 01:53
To: Ope...@li...
Subject: [users] troubles getting HA Assignment

Hi All,

We are using OpenSAF 5.2.0, and we are utilizing the OpenSAF Node Group
extension, which allows for Admin Commands to be issued in parallel to nodes
in the cluster.   The node group commands all work as expected, and we are
getting back a success code. However, the nodes never transition to a state
where they receive an HA assignment.   Below is the sequence of commands:


  1.  All Nodes in cluster are fully up and operational
  2.  Send Node Group Command to Lock all Nodes
     *   Success - Nodes and SUs achieve LOCK state
  3.  Send Node Group Command to Lock_Instantiate all Nodes
     *   Success - Nodes and SUs achieve LOCK_Instantiate state
  4.  Note - at this point, the cluster is completely down, as expected.
The only processes running are OpenSAF processes.
  5.  Send Node Group Command to UnLock_Instantiate all Nodes
     *   Success - Nodes and SUs achieve LOCK state
  6.  Send Node Group Command to UnLock all Nodes
     *   Success - Nodes and SUs achieve UnLock state
  7.  At this point, the system should be fully operational.  But instead we
have the following:
     *   Node States:

                                                    i.     Admin State =
unlocked

                                                   ii.     Operational State
= enabled

     *   SU States:

                                                    i.     Admin State =
unlocked

                                                   ii.     Operational State
= enabled

                                                 iii.     Presence State =
instantiated

                                                 iv.     READINESS STATE =
OUT OF SERVICE

                                                  v.     HA STATE = NONE
ASSIGNED

We cannot figure out why the Readiness State is OUT OF SERVICE.  It seems
like something has prevented the AMF from assigning an HA STATE to the SUs,
but it is not clear what this is.  Has anyone encountered a similar issue,
where HA STATE were not getting assigned?   Does anyone have a
recommendation on how to troubleshoot this issue?

Thanks

Jim


_______________________________________________
Opensaf-users mailing list
Ope...@li...
https://lists.sourceforge.net/lists/listinfo/opensaf-users