尝试安装Acumos时出错

Question

背景: VMware15.0 ubuntu16.04-64bit 32G RAM + 16Core CPU /etc/hosts: 192.168.79.130 localhost

执行此操作(并在出现提示时键入sudo密码)：

git clone https://gerrit.acumos.org/r/system-integration
apt-get -y update
apt-get -y install docker-ce=18.06.3~ce~3-0~ubuntu
if [[ "$(id -nG "$USER" | grep docker)" == "" ]]; then sudo usermod -aG docker $USER; fi
# Logged out and in again and verified that my user is in the docker group
cd system-integration/tools/
bash setup_k8s_stack.sh setup
cd
bash system-integration/AIO/setup_prereqs.sh k8s localhost $USER generic 2>&1 | tee aio_prep.log
# When "Prerequisites setup is complete" messages is displayed I continue with
cd system-integration/AIO
bash oneclick_deploy.sh 2>&1 | tee aio_deploy.log

部署失败，并显示以下错误消息：

+ c='-l component=192.168.79.130'
++ kubectl get deployment -n acumos -l app=cds -l component=192.168.79.130 -o json
++ jq -r '.items[0].metadata.name'
+ dep=null
++ cat /tmp/a72a447b-df96-4fec-98c9-bb99e447d00d
+ kubectl patch deployment -n acumos null --patch 'spec:
  template:
    spec:
      hostAliases:
      - ip: "192.168.79.130"
        hostnames:
        - "ubuntu"'
Error from server (NotFound): deployments.extensions "null" not found

打开文件：“system-integration/AIO/utils.sh”

  if [[ "$component" != "" ]]; then c="-l component=$component"; fi
  dep=$(kubectl get deployment -n $namespace -l app=$app $c -o json | jq -r ".items[0].metadata.name")
  kubectl patch deployment -n $namespace $dep --patch "$(cat $tmp)"

修改为：

  #if [[ "$component" != "" ]]; then c="-l component=$component"; fi
  dep=$(kubectl get deployment -n $namespace -l app=$app $c -o json | jq -r ".items[0].metadata.name")
  kubectl patch deployment -n $namespace filebeat --patch "$(cat $tmp)"

已解决此错误，但出现以下错误：

oneclick_deploy.sh setup_federation:233 (Tue Sep 17 02:32:31 PDT 2019) CDS API is not yet ready; waiting 10 seconds
+ t=300
+ sleep 10
++ curl -k -u ccds_client:187bbf19-40b9-45c8-9945-4903292d963d https://localhost/ccds/peer
++ grep -c numberOfElements
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100   203  100   203    0     0    957      0 --:--:-- --:--:-- --:--:--   962
+ [[ 0 -eq 0 ]]
+ [[ 300 -eq 300 ]]
+ fail 'CDS API is not ready after 300 seconds'
+ set +x

oneclick_deploy.sh fail:42 (Tue Sep 17 02:32:41 PDT 2019) CDS API is not ready after 300 seconds

pod cds出现错误: CrashLoopBackOff日志：

Events:
  Type     Reason     Age                  From               Message
  ----     ------     ----                 ----               -------
  Normal   Scheduled  13m                  default-scheduler  Successfully assigned acumos/cds-7474fccbc7-nm4jk to ubuntu
  Normal   Pulled     9m28s (x5 over 13m)  kubelet, ubuntu    Container image "nexus3.acumos.org:10002/common-dataservice:2.2.5" already present on machine
  Normal   Created    9m28s (x5 over 13m)  kubelet, ubuntu    Created container
  Normal   Started    9m28s (x5 over 13m)  kubelet, ubuntu    Started container
  Warning  BackOff    3m9s (x28 over 12m)  kubelet, ubuntu    Back-off restarting failed container

这个问题类似于前面的问题：Errors trying to install Acumos Boreas release，谢谢！

日志：(‘cds的补丁部署’)

...
oneclick_deploy.sh start_deployment:518 (Tue Sep 17 18:25:16 PDT 2019) Creating deployment cds
+ kubectl create -f deploy/cds-deployment.yaml
deployment.apps/cds created
+ get_host_ip_from_etc_hosts localhost
+ trap fail ERR
++ grep -v '^127\.'
++ awk '{print $1}'
++ grep -E '\slocalhost( |$)' /etc/hosts
+ HOST_IP='192.168.79.130
192.168.79.130'
+ [[ 192.168.79.130
192.168.79.130 != '' ]]
+ patch_deployment_with_host_alias acumos cds ubuntu 192.168.79.130 192.168.79.130
+ trap fail ERR
+ namespace=acumos
+ app=cds
+ name=ubuntu
+ ip=192.168.79.130
+ component=192.168.79.130
+ log 'Patch deployment for cds (192.168.79.130), to restart it with the changes'
+ setx=x
+ set +x

oneclick_deploy.sh patch_deployment_with_host_alias:448 (Tue Sep 17 18:25:17 PDT 2019) Patch deployment for cds (192.168.79.130), to restart it with the changes
++ uuidgen
+ tmp=/tmp/ef6cab06-ece6-436b-812c-1a00728bec01
+ cat
+ [[ 192.168.79.130 != '' ]]
+ c='-l component=192.168.79.130'
++ kubectl get deployment -n acumos -l app=cds -l component=192.168.79.130 -o json
...

Answer 1

两个建议：

(1)添加您自己的无密码sudo权限

sudo visudo
(add to the end of the file and save)
<your username>   ALL=(ALL:ALL) NOPASSWD:ALL

2) Re‘部署失败，错误消息如下：’，'-l component=192.168.79.130‘表示一个错误，导致patch_deployment_with_host_alias (在utils.sh中)认为指定了一个组件(参数5)。对此的具体调用来自cds应用程序的start_acumos_core_app，在一行中

patch_deployment_with_host_alias $ACUMOS_NAMESPACE $app $ACUMOS_MARIADB_HOST $HOST_IP  
or
patch_deployment_with_host_alias $ACUMOS_NAMESPACE $app $ACUMOS_HOST $HOST_IP

由于其中最多只有四个参数(截至master)，您是否可以提供更多的日志信息(返回到输出行：'Patch deployment for cds')。

除此之外，不需要更新utils.sh，我认为您所做的更改可能有副作用。所以我会撤销这些更改：-你用filebeat替换了$dep，这可能会让你进步，但它破坏了那个功能的目的(为组件添加主机别名-而不仅仅是filebeat -这些组件引用了DNS中无法解析的名称)。