Microsoft MVP성태의 닷넷 이야기
글쓴 사람
정성태 (techsharer at outlook.com)
홈페이지
첨부 파일
 

AKS에서 hpa에 따른 autoscale 기능이 동작하지 않는다면?

결과 먼저 말하자면, 원인을 밝혀내진 못했으니 그냥 참고삼아 보시면 되겠습니다. ^^




Azure에 생성한 AKS 서비스에서 autoscale 설정을 했는데,

$ kubectl autoscale deployment --max=10 net-razor31-sample --min=2

아무리 부하를 줘도 pod는 그대로 2개에 머물고 있습니다.

$ kubectl get pods
NAME                                  READY   STATUS    RESTARTS   AGE
net-razor31-sample-6b747bb886-ncvvw   1/1     Running   0          4m51s
net-razor31-sample-6b747bb886-sxkbj   1/1     Running   0          5m6s

hpa 설정을 보면 이상 없는 듯한데요,

$ kubectl get hpa
NAME                 REFERENCE                       TARGETS         MINPODS   MAXPODS   REPLICAS   AGE
net-razor31-sample   Deployment/net-razor31-sample   <unknown>/80%   2         10        2          6m50s

실제로는 이상이 있는 것입니다. ^^ 문제가 되는 부분은 "TARGETS"의 조건인데 여기에 "<unknown>"이 나오는 것은 현재 해당 시스템의 성능 metrics 데이터를 정상적으로 가져오지 못한다는 것을 의미합니다. 따라서 부하 측정을 할 수 없고 당연히 autoscale을 위한 기준을 잡지 못하는 것입니다.




기본적으로 (근래 버전의 k8s를 선택한 경우) AKS는 metrics-server 구성 요소를 이미 준비하고 있습니다.

C:\temp> kubectl get pod --all-namespaces
NAMESPACE     NAME                                  READY   STATUS    RESTARTS   AGE
default       net-razor31-sample-6b747bb886-mldk2   0/1     Pending   0          107s
kube-system   azure-ip-masq-agent-tmqtq             1/1     Running   0          6h
kube-system   coredns-845757d86-7bzwn               1/1     Running   0          17h
kube-system   coredns-845757d86-rljs8               1/1     Running   0          17h
kube-system   coredns-autoscaler-5f85dc856b-m9s57   1/1     Running   0          17h
kube-system   csi-azuredisk-node-ltfl4              3/3     Running   0          6h
kube-system   csi-azurefile-node-8xggf              3/3     Running   0          6h
kube-system   kube-proxy-p4zj6                      1/1     Running   0          17h
kube-system   metrics-server-6bc97b47f7-6tqms       1/1     Running   0          17h
kube-system   tunnelfront-78bc5d7569-fvlth          1/1     Running   0          16h

따라서 성능 메트릭 데이터를 수집하지 못할 이유가 없는데요, 이에 대한 문제를 describe hpa로 살펴볼 수 있습니다.

c:\temp> kubectl describe hpa net-razor31-sample
Name:                                                  net-razor31-sample
Namespace:                                             default
Labels:                                                <none>
Annotations:                                           <none>
CreationTimestamp:                                     Tue, 25 Jan 2022 15:44:52 +0900
Reference:                                             Deployment/net-razor31-sample
Metrics:                                               ( current / target )
  resource cpu on pods  (as a percentage of request):  <unknown> / 80%
Min replicas:                                          2
Max replicas:                                          10
Deployment pods:                                       2 current / 2 desired
Conditions:
  Type           Status  Reason                   Message
  ----           ------  ------                   -------
  AbleToScale    True    SucceededGetScale        the HPA controller was able to get the target's current scale
  ScalingActive  False   FailedGetResourceMetric  the HPA was unable to compute the replica count: failed to get cpu utilization: missing request for cpu
Events:
  Type     Reason                        Age   From                       Message
  ----     ------                        ----  ----                       -------
  Normal   SuccessfulRescale             33s   horizontal-pod-autoscaler  New size: 2; reason: Current number of replicas below Spec.MinReplicas
  Warning  FailedGetResourceMetric       18s   horizontal-pod-autoscaler  failed to get cpu utilization: unable to get metrics for resource cpu: no metrics returned from resource metrics API
  Warning  FailedComputeMetricsReplicas  18s   horizontal-pod-autoscaler  invalid metrics (1 invalid out of 1), first error is: failed to get cpu utilization: unable to get metrics for resource cpu: no metrics returned from resource metrics API
  Warning  FailedGetResourceMetric       3s    horizontal-pod-autoscaler  failed to get cpu utilization: missing request for cpu
  Warning  FailedComputeMetricsReplicas  3s    horizontal-pod-autoscaler  invalid metrics (1 invalid out of 1), first error is: failed to get cpu utilization: missing request for cpu

위의 결과에 보면, "failed to get cpu utilization: unable to get metrics for resource cpu: no metrics returned from resource metrics API"라는 메시지가 있는데요, 뭔가 해당 pod에서 CPU 자원을 구하는 코드가 동작하지 않는 듯합니다.

이상한 것은, top nodes나 top pods 옵션에서는 CPU 정보를 잘 가져온다는 점입니다.

c:\temp> kubectl top nodes
NAME                                CPU(cores)   CPU%   MEMORY(bytes)   MEMORY%
aks-agentpool-20726260-vmss000000   216m         11%    2073Mi          45%

c:\temp> kubectl top pods
NAME                                  CPU(cores)   MEMORY(bytes)
net-razor31-sample-574ffd4758-4zlxm   2m           52Mi
net-razor31-sample-574ffd4758-7szr8   1m           54Mi
sample-back-56bf7ff864-7hdzm          2m           14Mi
sample-front-65fd487fdf-jjhwh         1m           43Mi
sample-front-65fd487fdf-sjcbr         1m           43Mi

참... 이걸 뭐라고 설명해야 할지 모르겠군요. ^^;




웹 검색을 했지만, 딱히 이거라고 할 만한 오류 원인을 찾지 못했습니다. 그래서 어쩔 수 없이 AKS 클러스터를 새로 생성하는 것으로 해결을 했습니다.

단지, 이전 실습과 다르게 한 점이 있다면,

AKS - Azure Kubernetes Service 생성 및 SLO/SLA 변경 방법
; https://www.sysnet.pe.kr/2/0/12922

이번에는 클러스터 생성 시 ACR까지 함께 생성했다는 점입니다. 그 외에는 (쿠버네티스 버전을 포함한) 모든 옵션들이 이전 클러스터 생성과 다른 점이 없었습니다. 하지만 상식적으로 ACR이 성능 메트릭스를 가져오는 것에 영향을 미쳤을 것 같지는 않고, 테스트 플랫폼이라 이거저거 실습하느라 건드린 무엇인가가... 문제였을 듯합니다.




[이 글에 대해서 여러분들과 의견을 공유하고 싶습니다. 틀리거나 미흡한 부분 또는 의문 사항이 있으시면 언제든 댓글 남겨주십시오.]







[최초 등록일: ]
[최종 수정일: 1/26/2022]

Creative Commons License
이 저작물은 크리에이티브 커먼즈 코리아 저작자표시-비영리-변경금지 2.0 대한민국 라이센스에 따라 이용하실 수 있습니다.
by SeongTae Jeong, mailto:techsharer at outlook.com

비밀번호

댓글 작성자
 




1  2  3  4  5  [6]  7  8  9  10  11  12  13  14  15  ...
NoWriterDateCnt.TitleFile(s)
13793정성태10/28/20245149C/C++: 183. C++ - 윈도우에서 한글(및 유니코드)을 포함한 콘솔 프로그램을 컴파일 및 실행하는 방법
13792정성태10/27/20244633Linux: 99. Linux - 프로세스의 실행 파일 경로 확인
13791정성태10/27/20244901Windows: 267. Win32 API의 A(ANSI) 버전은 DBCS를 사용할까요?파일 다운로드1
13790정성태10/27/20244621Linux: 98. Ubuntu 22.04 - 리눅스 커널 빌드 및 업그레이드
13789정성태10/27/20244914Linux: 97. menuconfig에 CONFIG_DEBUG_INFO_BTF, CONFIG_DEBUG_INFO_BTF_MODULES 옵션이 없는 경우
13788정성태10/26/20244459Linux: 96. eBPF (bpf2go) - fentry, fexit를 이용한 트레이스
13787정성태10/26/20244973개발 환경 구성: 730. github - Linux 커널 repo를 윈도우 환경에서 git clone하는 방법 [1]
13786정성태10/26/20245218Windows: 266. Windows - 대소문자 구분이 가능한 파일 시스템
13785정성태10/23/20244985C/C++: 182. 윈도우가 운영하는 2개의 Code Page파일 다운로드1
13784정성태10/23/20245248Linux: 95. eBPF - kprobe를 이용한 트레이스
13783정성태10/23/20244860Linux: 94. eBPF - vmlinux.h 헤더 포함하는 방법 (bpf2go에서 사용)
13782정성태10/23/20244621Linux: 93. Ubuntu 22.04 - 커널 이미지로부터 커널 함수 역어셈블
13781정성태10/22/20244801오류 유형: 930. WSL + eBPF: modprobe: FATAL: Module kheaders not found in directory
13780정성태10/22/20245550Linux: 92. WSL 2 - 커널 이미지로부터 커널 함수 역어셈블
13779정성태10/22/20244845개발 환경 구성: 729. WSL 2 - Mariner VM 커널 이미지 업데이트 방법
13778정성태10/21/20245671C/C++: 181. C/C++ - 소스코드 파일의 인코딩, 바이너리 모듈 상태의 인코딩
13777정성태10/20/20244952Windows: 265. Win32 API의 W(유니코드) 버전은 UCS-2일까요? UTF-16 인코딩일까요?
13776정성태10/19/20245269C/C++: 180. C++ - 고수준 FILE I/O 함수에서의 Unicode stream 모드(_O_WTEXT, _O_U16TEXT, _O_U8TEXT)파일 다운로드1
13775정성태10/19/20245492개발 환경 구성: 728. 윈도우 환경의 개발자를 위한 UTF-8 환경 설정
13774정성태10/18/20245196Linux: 91. Container 환경에서 출력하는 eBPF bpf_get_current_pid_tgid의 pid가 존재하지 않는 이유
13773정성태10/18/20244883Linux: 90. pid 네임스페이스 구성으로 본 WSL 2 + docker-desktop
13772정성태10/17/20245160Linux: 89. pid 네임스페이스 구성으로 본 WSL 2 배포본의 계층 관계
13771정성태10/17/20245066Linux: 88. WSL 2 리눅스 배포본 내에서의 pid 네임스페이스 구성
13770정성태10/17/20245337Linux: 87. ps + grep 조합에서 grep 명령어를 사용한 프로세스를 출력에서 제거하는 방법
13769정성태10/15/20246114Linux: 86. Golang + bpf2go를 사용한 eBPF 기본 예제파일 다운로드1
13768정성태10/15/20245396C/C++: 179. C++ - _O_WTEXT, _O_U16TEXT, _O_U8TEXT의 Unicode stream 모드파일 다운로드2
1  2  3  4  5  [6]  7  8  9  10  11  12  13  14  15  ...