When the SSP UI is unavailable but the SSPI CLI is accessible, it is still possible to collect a support bundle using the CLI.
If the cluster-api pod is running, the standard support bundle collection procedure can be followed as documented.
However, when the cluster-api pod is down, collecting a full support bundle becomes challenging.
This article describes a workaround to manually collect critical logs when the cluster-api pod is unavailable.
cluster-api pod downThe cluster-api pod is the primary component responsible for coordinating interactions between SSP services and pods.
When this pod is down, normal support bundle generation is not possible.
In this scenario, the fluentd pod continues to run and aggregates logs from across the cluster. These logs can be manually collected and packaged as a workaround support bundle.
From the SSPI CLI, confirm that the fluentd pod is in a Running state:
Inside the fluentd pod, confirm logs are present:
Example output includes:
audit_log.log
kube-system/
nsxi-platform/
n1ssp-controller-*
n1ssp-md-*
Other component-specific logs
Navigate further as needed:
td-agent@fluentd-0:/$ ls /opt/bitnami/fluentd/logs/buffers/nsxi-platform/
Use kubectl cp to copy the buffered logs to the SSPI node:
Note:
Warnings such asfile changed as we read itare expected because logs are actively being written. These warnings can be safely ignored.
Confirm that the directory structure and log files are present.
Compress the collected logs for easier handling and transfer:
Use scp or sftp from your local system to retrieve the bundle:
scp sysadmin@n1sspi:/home/sysadmin/ssp_supportbundle.tar.gz .