Issue with SUSE Trento Agent on SLES 15 SP3 for SAP Applications on IBM Power

[williandrade82]

We encountered an issue with the SUSE Trento Agent on a SLES 15 SP3 (for SAP Application) deployed on an IBM Power 9’s LPAR. The agent relies on dmidecode to read low-level system information, which is not available on IBM Power servers. This results in errors during the agent's execution.

Environment:

• Operating System: SLES 15 SP3 (for SAP Applications)
• Hardware: IBM Power 9’s LPAR
• Application: SAP HANA database
• Issue Description: The SUSE Trento Agent logs errors related to the absence of the dmidecode package, which is not available on IBM Power servers. Below is an example of the errors logged (grep from journalctl of the trento-agent unit):

2025-02-28T10:37:01.779541-03:00 anonymized-hostname trento-agent[34106]: time="2025-02-28 10:37:01" level=info msg="Identifying if the system is running in a cloud environment..."
2025-02-28T10:37:01.779700-03:00 anonymized-hostname trento-agent[34106]: time="2025-02-28 10:37:01" level=error msg="Error while running discovery 'cloud_discovery': exec: \"dmidecode\": executable file not found in $PATH"
2025-02-28T10:37:01.779786-03:00 anonymized-hostname trento-agent[34106]: time="2025-02-28 10:37:01" level=info msg="cloud_discovery discovery tick output: Error while running discovery 'cloud_discovery': exec: \"dmidecode\": executable file not found in $PATH"
2025-02-28T10:37:01.811145-03:00 anonymized-hostname trento-agent[34106]: time="2025-02-28 10:37:01" level=error msg="Unable to convert CPU socket count: strconv.Atoi: parsing \"\": invalid syntax"
2025-02-28T10:37:01.822480-03:00 anonymized-hostname trento-agent[34106]: time="2025-02-28 10:37:01" level=info msg="ha_cluster_discovery discovery tick output: No HA cluster discovered on this host"
2025-02-28T10:37:01.835556-03:00 anonymized-hostname trento-agent[34106]: time="2025-02-28 10:37:01" level=info msg="host_discovery discovery tick output: Host with name: anonymized-hostname successfully discovered"
2025-02-28T10:37:02.284597-03:00 anonymized-hostname trento-agent[34106]: time="2025-02-28 10:37:02" level=info msg="New SAP system installation found: CSD"
2025-02-28T10:37:02.284699-03:00 anonymized-hostname trento-agent[34106]: time="2025-02-28 10:37:02" level=info msg="New SAP system installation found: DAA"
2025-02-28T10:37:02.284810-03:00 anonymized-hostname trento-agent[34106]: time="2025-02-28 10:37:02" level=info msg="New SAP instance installation found: ASCS21"
2025-02-28T10:37:02.284875-03:00 anonymized-hostname trento-agent[34106]: time="2025-02-28 10:37:02" level=info msg="New SAP instance installation found: D20"
2025-02-28T10:37:02.724366-03:00 anonymized-hostname trento-agent[34106]: time="2025-02-28 10:37:02" level=info msg="New SAP instance installation found: SMDA98"
2025-02-28T10:37:02.768402-03:00 anonymized-hostname trento-agent[34106]: time="2025-02-28 10:37:02" level=info msg="sap_system_discovery discovery tick output: SAP system(s) with ID: CSD,DAA discovered"

Impact: The absence of dmidecode on IBM Power servers causes the SUSE Trento Agent to fail in gathering necessary system information, leading to incomplete or inaccurate system discovery and monitoring.

Suggested Solution: To address this issue, the SUSE Trento Agent should be updated to use alternative methods for gathering system information on IBM Power servers, such as utilizing the device tree structure (/proc/device-tree/) or the lshw command.

[stefanotorresi]

Hey @williandrade82 thanks for letting us know, we are aware of the issue and we'll look into it ASAP.
Have you reported this via SUSE Customer Center as well?

[williandrade82]

Thank you for your response. I haven't opened a ticket with the SUSE Customer Center because the issue doesn't seem to be related to the operating system itself. As indicated in the application logs, the application agent is attempting to use a resource (dmidecode) that is not available on IBM Power architecture. I believe the solution would be to update the agent to use alternative methods for gathering system information, such as the device tree structure (/proc/device-tree/) or the lshw command.

However, if you consider it important to report this issue through the SUSE Customer Center, please let me know, and I will do so.

Thank you for your attention, and I'm available for any further clarification.

[stefanotorresi]

However, if you consider it important to report this issue through the SUSE Customer Center, please let me know, and I will do so.

nono it's okay, it was only to cross-reference another similar report. For the record, Trento is supported as part of any SLES4SAP subscription, so if you want to escalate the issue to our enterprise support, feel free to do so via SCC.

Thanks for the suggestion on the alternative methods to gather the same info, we'll look into it over the coming weeks.

[stefanotorresi]

@williandrade82 other than the errors logged by the agent systemd unit, could you please give us more details on what is the unexpected observed behaviour?