Weekly Notes

Here are my weekly notes:

Flow Connector

If you are using the VPC, you can track connections between your subnets and your VPC using Flow Connector.

❯ find . -name “*.gz” -exec gunzip {} \;

❯ grep -Rh 192.168.200.10 | jq -r ‘.flow_logs[] | select(.action == “rejected”) | “\(.initiator_ip),\(.target_ip),\(.target_port)”‘ | sort -u | grep 192.168.200.10

10.245.0.5,192.168.200.10,36416,2023-08-08T14:31:32Z

10.245.0.5,192.168.200.10,36430,2023-08-08T14:31:32Z

10.245.0.5,192.168.200.10,58894,2023-08-08T14:31:32Z

10.245.1.5,192.168.200.10,10250,2023-08-08T14:31:41Z

10.245.1.5,192.168.200.10,10250,2023-08-08T14:31:42Z

10.245.1.5,192.168.200.10,9100,2023-08-08T14:31:32Z

10.245.129.4,192.168.200.10,43524,2023-08-08T14:31:32Z

10.245.64.4,192.168.200.10,10250,2023-08-08T14:31:32Z

10.245.64.4,192.168.200.10,10250,2023-08-08T14:31:42Z

10.245.64.4,192.168.200.10,9100,2023-08-08T14:31:42Z

10.245.64.4,192.168.200.10,9537,2023-08-08T14:50:36Z

Image Pruner Reports Error….

You can check the image-registry status on the cluster operator.

❯ oc get co image-registry
image-registry                             4.14.0-ec.4   True        False         True       3d14h   ImagePrunerDegraded: Job has reached the specified backoff limit

The cronjob probably failed, so we can check that it exists.

❯ oc get cronjob -n openshift-image-registry
NAME           SCHEDULE    SUSPEND   ACTIVE   LAST SCHEDULE   AGE
image-pruner   0 0 * * *   False     0        16h             3d15h

We can run a one-off to clear the status above.

❯ oc create job --from=cronjob/image-pruner one-off-image-pruner -n openshift-image-registry
job.batch/one-off-image-pruner created

Then your image-registry should be a-ok.

Ref: https://gist.github.com/ryderdamen/73ff9f93cd61d5dd45a0c50032e3ae03