One extra note: it's not necessary `pause` process, or `cephcsi` pod like in the example above. The pod/process selection also looks random, or at least I couldn't find any logic behind which one get's "UNKNOWN".
On this very node it's
# ctr --namespace k8s.io task ls|grep -i unknown 5f78e0cb957de97fd8465cc42c842bdd764d981ca7a903a2515bbc6bb06796a9 0 UNKNOWN 4e063ef0c8f768dbf34cf7a179bca5cc98a04fa7e00b29d20c17d3031d409f86 0 UNKNOWN af070f16c1f0ff22eb16661e787e85db3810727909abd23d69a6a43578c1dced 0 UNKNOWN 2f5a8376b476809b1696b140ca87f91422113bb16b27a8174437cc63b48e259a 0 UNKNOWN
where as one may see all the tasks belong the same pod.
One extra note: it's not necessary `pause` process, or `cephcsi` pod like in the example above. The pod/process selection also looks random, or at least I couldn't find any logic behind which one get's "UNKNOWN".
On this very node it's
# ctr --namespace k8s.io task ls|grep -i unknown fd8465cc42c842b dd764d981ca7a90 3a2515bbc6bb067 96a9 0 UNKNOWN bf34cf7a179bca5 cc98a04fa7e00b2 9d20c17d3031d40 9f86 0 UNKNOWN 2eb16661e787e85 db3810727909abd 23d69a6a43578c1 dced 0 UNKNOWN b1696b140ca87f9 1422113bb16b27a 8174437cc63b48e 259a 0 UNKNOWN
5f78e0cb957de97
4e063ef0c8f768d
af070f16c1f0ff2
2f5a8376b476809
where as one may see all the tasks belong the same pod.