Red HatでOpenShiftのサポートをしているid:nekopです。OpenShift 全部俺 Advent Calendar 2018 - Qiitaの12日目のエントリです。
デプロイメントの状況をや問題を確認するにはEventログとPodログを見ることになります。
OpenShift 3.11でoc get eventsすると以下のような出力となります。
$ oc run sleep --image=registry.access.redhat.com/rhel-minimal --command -- tail -f /dev/null
$ oc get events
LAST SEEN FIRST SEEN COUNT NAME KIND SUBOBJECT TYPE REASON SOURCE MESSAGE
1m 1m 1 sleep-1-deploy.15701ca29ab92fac Pod Normal Scheduled default-scheduler Successfully assigned test-events/sleep-1-deploy to node-1.example.com
1m 1m 1 sleep.15701ca29688a78d DeploymentConfig Normal DeploymentCreated deploymentconfig-controller Created new replication controller "sleep-1" for version 1
58s 58s 1 sleep-1.15701ca69bfcbb7c ReplicationController Normal SuccessfulCreate replication-controller Created pod: sleep-1-tjslb
58s 58s 1 sleep-1-tjslb.15701ca69d04b8bf Pod Normal Scheduled default-scheduler Successfully assigned test-events/sleep-1-tjslb to node-2.kaizen.lab.rdu2.cee.redhat.com
46s 46s 1 sleep-1-deploy.15701ca97ebc7af2 Pod spec.containers{deployment} Normal Pulled kubelet, node-1.example.com Container image "registry.redhat.io/openshift3/ose-deployer:v3.11.16" already present on machine
43s 43s 1 sleep-1-deploy.15701caa38a773dc Pod spec.containers{deployment} Normal Created kubelet, node-1.example.com Created container
42s 42s 1 sleep-1-deploy.15701caa7be825e2 Pod spec.containers{deployment} Normal Started kubelet, node-1.example.com Started container
21s 21s 1 sleep-1-tjslb.15701caf3db14837 Pod spec.containers{sleep} Normal Pulling kubelet, node-2.kaizen.lab.rdu2.cee.redhat.com pulling image "registry.access.redhat.com/rhel-minimal"
19s 19s 1 sleep-1-tjslb.15701cafc71c1b54 Pod spec.containers{sleep} Normal Pulled kubelet, node-2.kaizen.lab.rdu2.cee.redhat.com Successfully pulled image "registry.access.redhat.com/rhel-minimal"
18s 18s 1 sleep-1-tjslb.15701cb01467b985 Pod spec.containers{sleep} Normal Created kubelet, node-2.kaizen.lab.rdu2.cee.redhat.com Created container
17s 17s 1 sleep-1-tjslb.15701cb033d15221 Pod spec.containers{sleep} Normal Started kubelet, node-2.kaizen.lab.rdu2.cee.redhat.com Started container
ちなみにOpenShift 3.11以前でoc get eventsすると、KINDでソートされていました。Eventは時間順で表示される、というバイアスがあるのでちょっと不自然に感じましたが、3.11で変更されたようです。
$ oc get events
LAST SEEN FIRST SEEN COUNT NAME KIND SUBOBJECT TYPE REASON SOURCE MESSAGE
52s 52s 1 sleep-1-9lsfm.15701caef94c7362 Pod Normal Scheduled default-scheduler Successfully assigned sleep-1-9lsfm to s39.example.com
52s 52s 1 sleep-1-9lsfm.15701caf041ef4b8 Pod Normal SuccessfulMountVolume kubelet, s39.example.com MountVolume.SetUp succeeded for volume "default-token-dftpl"
50s 50s 1 sleep-1-9lsfm.15701caf8bc51822 Pod spec.containers{sleep} Normal Pulling kubelet, s39.example.com pulling image "registry.access.redhat.com/rhel-minimal"
48s 48s 1 sleep-1-9lsfm.15701caff5b87264 Pod spec.containers{sleep} Normal Pulled kubelet, s39.example.com Successfully pulled image "registry.access.redhat.com/rhel-minimal"
48s 48s 1 sleep-1-9lsfm.15701caffb4f9612 Pod spec.containers{sleep} Normal Created kubelet, s39.example.com Created container
48s 48s 1 sleep-1-9lsfm.15701cb00ce7b8d5 Pod spec.containers{sleep} Normal Started kubelet, s39.example.com Started container
58s 58s 1 sleep-1-deploy.15701cad97a0173b Pod Normal Scheduled default-scheduler Successfully assigned sleep-1-deploy to s39.example.com
58s 58s 1 sleep-1-deploy.15701cada6db286d Pod Normal SuccessfulMountVolume kubelet, s39.example.com MountVolume.SetUp succeeded for volume "deployer-token-756rx"
56s 56s 1 sleep-1-deploy.15701cae3b3dba8f Pod spec.containers{deployment} Normal Pulled kubelet, s39.example.com Container image "openshift3/ose-deployer:v3.9.40" already present on machine
56s 56s 1 sleep-1-deploy.15701cae3f6b9288 Pod spec.containers{deployment} Normal Created kubelet, s39.example.com Created container
55s 55s 1 sleep-1-deploy.15701cae623efbea Pod spec.containers{deployment} Normal Started kubelet, s39.example.com Started container
46s 46s 1 sleep-1-deploy.15701cb081a839c3 Pod spec.containers{deployment} Normal Killing kubelet, s39.example.com Killing container with id docker://deployment:Need to kill Pod
53s 53s 1 sleep-1.15701caef7fcf71e ReplicationController Normal SuccessfulCreate replication-controller Created pod: sleep-1-9lsfm
58s 58s 1 sleep.15701cad9505cfec DeploymentConfig Normal DeploymentCreated deploymentconfig-controller Created new replication controller "sleep-1" for version 1
また、3.10ではoc get events -wとしたときに以下のようにAbsolute timestampで表示されていたのですが、3.11では相対時間表示のままになっています。
$ oc get events -w LAST SEEN FIRST SEEN COUNT NAME KIND SUBOBJECT TYPE REASON SOURCE MESSAGE 2018-12-14 16:51:13 +0900 JST 2018-12-14 16:51:13 +0900 JST 1 sleep-1-cwfjr.157023a7312a611c Pod Normal Scheduled default-scheduler Successfully assigned sleep-1-cwfjr to s39.usersys.redhat.com
エンジニアリングに確認したところ意図的に変更を行ったとの回答だったので、実装を調べてみます。関連するPRは以下です。
- https://github.com/kubernetes/kubernetes/pull/66643
- https://github.com/kubernetes/kubernetes/pull/66643/commits/2f275b72b2248bb7c522327115acb17223b8387b
- https://github.com/kubernetes/kubernetes/commit/a3a430ce01c30a9773015dc78bc78dc8953a421c
oc get events -wの出力はAbsoluteTimestampsというフラグで切り替えています。1.10では以下のコードとなっていました。
$ cd kubernetes # https://github.com/kubernetes/kubernetes/
$ git co release-1.10
$ git grep AbsoluteTimestamps
pkg/kubectl/cmd/util/printing.go: AbsoluteTimestamps: isWatch(cmd),
pkg/printers/interface.go: AbsoluteTimestamps bool
pkg/printers/internalversion/printers.go: if options.AbsoluteTimestamps {
1.11では以下になっています。
$ git co release-1.11
$ git grep AbsoluteTimestamps
pkg/kubectl/cmd/get/humanreadable_flags.go: AbsoluteTimestamps bool
pkg/kubectl/cmd/get/humanreadable_flags.go: AbsoluteTimestamps: false,
pkg/printers/interface.go: AbsoluteTimestamps bool
pkg/printers/internalversion/printers.go: if options.AbsoluteTimestamps {
1.10ではwatchではAbsoluteTimestampsをtrueにしますが、1.11ではfalseに設定するコードしかないので実質false固定となっています。
ちなみに3.11での表示順ですが、3.11以前のようにeventのソートは行っておらず、apiserverからのnatural orderに変更された結果、より見た目的に自然な順序になっているようです。
