[ClusterLabs] Are use of custom meta-attributes not handled by Pacemaker valid?

[alexey at pavlyuts.ru]

Hi All,

 

We have a project where Pacemaker is under the hood and all the high-level
business-logic managed by external custom software, mostly by reading and
writing cib XML with cibadmin binary.

 

I need to store some extra properties of the managed resources for these
business-logic handlers. Pacemaker itself need no this data.

 

I found that if I add custom nvpair element in primitive/meta_attributes to
cib xml and store it with cibadmin -replace it seems to be accepted and
stored in CIB database without problem. And reading back well keeping this
"custom meta-attribute".

 

Example:

      <primitive id="vlan-ens18_10-ens18.10" class="ocf" type="iface-vlan"
provider="heartbeat">

        <meta_attributes id="vlan-ens18_10-ens18.10-meta_attributes">

          <nvpair id="vlan-ens18_10-ens18.10-meta_attributes-custom"
name="custom" value="true"/>

        </meta_attributes>

        <instance_attributes
id="vlan-ens18_10-ens18.10-instance_attributes">

          <nvpair id="vlan-ens18_10-ens18.10-instance_attributes-vlan_id"
name="vlan_id" value="10"/>

          <nvpair
id="vlan-ens18_10-ens18.10-instance_attributes-vlan_interface"
name="vlan_interface" value="ens18"/>

          <nvpair id="vlan-ens18_10-ens18.10-instance_attributes-vlan_name"
name="vlan_name" value="ens18.10"/>

        </instance_attributes>

        <operations>

          <op name="monitor" interval="10s" timeout="20s"
id="vlan-ens18_10-ens18.10-monitor-interval-10s"/>

          <op name="start" interval="0s" timeout="30s"
id="vlan-ens18_10-ens18.10-start-interval-0s"/>

          <op name="stop" interval="0s" timeout="20s"
id="vlan-ens18_10-ens18.10-stop-interval-0s"/>

        </operations>

      </primitive>

 

 

But the question is: how valid this action is? For now, for future version?
Is there a chance that some of Pacekeeper internals or utilities wipe it up
or break on it?

 

I will be very grateful for any information on it, thank you in advance!

 

Sincerely,

 

Alex

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/users/attachments/20250227/8c58f104/attachment.htm>