One more reason that it should be done by CLM could be: AMF cluster can be smaller than CLM cluster. So to reset all the nodes only CLM can do it.
What is the use case of rebooting the cluster through AMF/CLM?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Agree CLM may be better suited, but this is a small change to AMF, and can be used for other use cases as well, e.g. after headless and if data inconsistency is detected a cluster reboot is needed.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Add support for cluster reboot via command immadm -a safClmService -o 100 safClmService. Clmsv then sends a reboot request message to each clm nodeagent. The clm nodeagents will call shutdown –r for a safe reboot.
Last edit: Hans Nordebäck 2016-09-23
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
As per CLM information model, This operation should be introduced as an admin operation on the object of SaClmCluster class.
i.e. immadm -o <new_operId> safCluster=myClmCluster
--- old+++ new@@ -1 +1 @@-Add support for cluster reboot via command immadm -a safAmfService -o 100 safAmfService+Add support for cluster reboot via command immadm -o 1 safCluster=myClmCluster
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
--- old+++ new@@ -1 +1 @@-Add support for cluster reboot via command immadm -o 1 safCluster=myClmCluster+Add support for cluster reboot via command immadm -o 4 safCluster=myClmCluster
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Possibly this is functionality should be in CLM? I.e. you send an admin-op to the CLM cluster object, to order a cluster reboot?
One more reason that it should be done by CLM could be: AMF cluster can be smaller than CLM cluster. So to reset all the nodes only CLM can do it.
What is the use case of rebooting the cluster through AMF/CLM?
Agree CLM may be better suited, but this is a small change to AMF, and can be used for other use cases as well, e.g. after headless and if data inconsistency is detected a cluster reboot is needed.
One use case is upgrade with a following cluster reboot.
Add support for cluster reboot via command immadm -a safClmService -o 100 safClmService. Clmsv then sends a reboot request message to each clm nodeagent. The clm nodeagents will call shutdown –r for a safe reboot.
Last edit: Hans Nordebäck 2016-09-23
As per CLM information model, This operation should be introduced as an admin operation on the object of SaClmCluster class.
i.e. immadm -o <new_operId> safCluster=myClmCluster
B.T.W i just created a ticket in AMF https://sourceforge.net/p/opensaf/tickets/2065/
Diff:
Diff:
changeset: 8183:f0b59b8c475c
tag: tip
parent: 8180:b4b4202688d8
user: Hans Nordeback hans.nordeback@ericsson.com
date: Thu Sep 29 15:50:07 2016 +0200
files: osaf/libs/common/clmsv/include/clmsv_msg.h osaf/libs/core/common/include/osaf_utility.h osaf/libs/core/common/osaf_utility.c osaf/libs/saf/include/saClm.h osaf/services/saf/clmsv/clms/clms.h osaf/services/saf/clmsv/clms/clms_imm.c osaf/services/saf/clmsv/clms/clms_mds.c osaf/services/saf/clmsv/clms/clms_util.c osaf/services/saf/clmsv/nodeagent/main.c scripts/opensaf_reboot
description:
clm: add support for cluster reboot V4 [#2053]
Admin command to request cluster reboot:
immadm -o 4 safCluster=myClmCluster
Related
Tickets:
#2053