QuestionQ33

Troubleshoot

Map the management-plane impact when Pensando Policy and Services Manager (PSM) is deployed in high availability (HA) and an HA issue occurs.

Drag & Drop
Configuration changes are accepted by the PSM and applied to DSSes Metrics and logs are regularly collected.
Configuration changes are accepted by the PSM but will not take effect until the DSS and PSM can communicate. Metrics and logs will be buffered at the DSS as local memory availability permits.
The PSM does not allow new user logins and does not accept configuration changes, to avoid a “split-brain” scenario. Metrics and logs are not collected by the PSM but will be buffered at the DSS subject to local storage availability.
Explanation

A single PSM-node failure retains HA-cluster quorum, so configuration delivery and telemetry collection continue. A DSS-to-PSM communication outage delays delivery to that DSS and causes it to buffer telemetry locally. Failure of two PSM nodes removes quorum; PSM prevents new logins and configuration changes to avoid split brain, while DSSs buffer telemetry subject to local storage capacity.

Community Discussion

No comments yet. Be the first to start the discussion!