When using SAPHanaSR-showAttr in a Pacemaker cluster that has both: * a SAP HANA controller resource (SAPHanaController), and

• an additional multi-state firewall resource that also uses master-rsc_* node attributes,

the secondary controller node score is reported incorrectly.

crm_mon -A -D -1 shows the correct scores (150 for the promoted master controller, 100 for the healthy demoted secondary), but SAPHanaSR-showAttr displays 5 for the healthy secondary instead of 100.

This appears to be caused by pattern matching in saphana_sr_tools.py that matches all master-rsc_SID_* attributes, including those created by the firewall resource, and then treating them as controller scores.
 
Current State:
Cluster configuration (simplified): * Controller resource name:

•  rsc_SAPHanaController_A23_HDB00 (type SAPHanaController, class ocf, provider heartbeat)
• Additional firewall resource (multi-state), used to avoid split-brain/racing conditions:
•  rsc_firewall_A23_HDB00
• The firewall resource is also promotable and sets node attributes of the form:
•  master-rsc_firewall_A23_HDB00

Node attributes example (for node syz1bab3dbn01):
 

Node Attributes:
  * Node: syz1bab3dbn01:
    * hana_a23_clone_state              : DEMOTED
    * hana_a23_roles                    : master1:master:worker:master
    * hana_a23_site                     : AB3AZ1
    * hana_a23_sra                      : -
    * hana_a23_srah                     : -
    * hana_a23_version                  : 2.00.079.02
    * hana_a23_vhost                    : syz1bab3dbn01
    * master-rsc_SAPHanaController_A23_HDB00    : 100
    * master-rsc_firewall_A23_HDB00             : 5

 
Behaviour of the tools: * crm_mon -A -D -1

• Shows correct controller scores:

• • • 150 for the healthy/promoted master controller

• • 100 for the healthy demoted secondary

• SAPHanaSR-showAttr

• • Shows 150 for the promoted master/controller (correct)
  • Shows only 5 for the healthy secondary (incorrect – should be 100)

Suspected root cause is in saphana_sr_tools.py (lines 79–81 in upstream at the time of writing):

# TODO: Do we need to check, if the master-attribute belongs to the promotable clone for this SID?
match_obj = re.search(f"master.rsc.*_{sid_uc}_.*", column_name)   # (4)
if match_obj is not None:
    column_name = 'score'
return column_name 

This code is imported and used by SAPHanaSR-showAttr (import on line 28).
 Because the regex is very broad, both of these attributes match: * master-rsc_SAPHanaController_A23_HDB00

• master-rsc_firewall_A23_HDB00

so the firewall’s 5 is effectively being used where the controller’s 100 should be reported.
 
Expected Outcome: * SAPHanaSR-showAttr should only interpret controller-related node attributes as the score for the SAP HANA controller resource, and ignore unrelated master-rsc_* attributes from other promotable resources (like the firewall).

• For the example above, SAPHanaSR-showAttr should:
  • Report 150 for the promoted master/controller
  • Report 100 for the healthy secondary controller
  • Not be affected by the master-rsc_firewall_A23_HDB00 : 5 attribute

 
Additional Notes / Possible Direction for Fix
A naïve workaround would be to hard-code the controller resource pattern, for example:
 

match_obj = re.search(f"master-rsc_SAPHanaController_{sid_uc}_.*", column_name) 

However, a better long-term fix would be to: * Find the name of the resource correlating with the LRM resource list such as:

•  <lrm_resource id="rsc_SAPHanaController_A23_HDB00" type="SAPHanaController" class="ocf" provider="heartbeat">), and use a specific resource name without hardcoding the naming convention