Logging

1. Introduction

• useful for debugging

• to see what's going on inside the containers

2. Creating a pod

• that writes to stdout and stderr

hadoop@k8s-00:~$ kubectl create -f https://raw.githubusercontent.com/mhausenblas/kbe/master/specs/logging/pod.yaml
pod "logme" created

3. To view the 5 most recent logs use

hadoop@k8s-00:~$ kubectl logs --tail=5 logme -c gen
Mon Jun 18 05:42:04 UTC 2018
Mon Jun 18 05:42:05 UTC 2018
Mon Jun 18 05:42:05 UTC 2018
Mon Jun 18 05:42:06 UTC 2018
Mon Jun 18 05:42:06 UTC 2018

where --tail=n specifies the n recent logs

• --since=n secs option, which is used to see the log details since the last n seconds

hadoop@k8s-00:~$ kubectl logs -f --since=10s logme -c gen
Mon Jun 18 05:45:59 UTC 2018
Mon Jun 18 05:45:59 UTC 2018
Mon Jun 18 05:46:00 UTC 2018
Mon Jun 18 05:46:00 UTC 2018
Mon Jun 18 05:46:01 UTC 2018
Mon Jun 18 05:46:01 UTC 2018
Mon Jun 18 05:46:02 UTC 2018

• to view the logs on the containers that have already completed their tasks, we can do follows

• create a pod oneshot

hadoop@k8s-00:~$ kubectl create -f https://raw.githubusercontent.com/mhausenblas/kbe/master/specs/logging/oneshotpod.yaml
pod "oneshot" created

• we can see that the pod has completed it's task

hadoop@k8s-00:~$ kubectl get pods
NAME      READY     STATUS             RESTARTS   AGE
badpod    0/1       CrashLoopBackOff   15         39m
hc        1/1       Running            0          45m
logme     1/1       Running            0          8m
oneshot   0/1       Completed          3          51s
ready     1/1       Running            0          33m

• still we can get the logs of the completed tasks as follows

hadoop@k8s-00:~$ kubectl logs -p oneshot -c gen
9
8
7
6
5
4
3
2
1