关于竹类连续部署的openshift cli问题

Question

我正在使用竹为连续Deployment.Now，如果运行一个竹计划(常规)，我的脚本运行完美。它可以获得路径(竹上的环境变量)上的oc (openshift命令)，例如下面的命令通过竹计划执行得很好:-

# include openshift in path
PATH=$PATH:/data/bamboo/openshift

/bin/sh ./portal/deploy.sh --build-number=${bamboo.buildNumber}

但是，如果我创建了一个“Deployment”，并尝试运行类似的命令，我总是会得到oc: not找到的错误

完全错误日志：-

26-Jul-2019 20:21:27    -> Getting present working dir...
26-Jul-2019 20:21:27    /opt/bamboo-agent-root/bamboo-agent-home/xml-data/build-dir/518946826-518455357
26-Jul-2019 20:21:27    -> Deployment started...
26-Jul-2019 20:21:27    -> Bringing down the pod(s)...
26-Jul-2019 20:21:27    -> Recreating the pod(s)...
26-Jul-2019 20:21:27    -> Deployment completed on TST environment
26-Jul-2019 20:21:27    /opt/bamboo-agent-root/bamboo-agent-home/temp/518946826-518455357-517442402-ScriptBuildTask-7080740712063035315.sh: 6: /opt/bamboo-agent-root/bamboo-agent-home/temp/518946826-518455357-517442402-ScriptBuildTask-7080740712063035315.sh: oc: not found
26-Jul-2019 20:21:27    /opt/bamboo-agent-root/bamboo-agent-home/temp/518946826-518455357-517442402-ScriptBuildTask-7080740712063035315.sh: 10: /opt/bamboo-agent-root/bamboo-agent-home/temp/518946826-518455357-517442402-ScriptBuildTask-7080740712063035315.sh: oc: not found
26-Jul-2019 20:21:27    /opt/bamboo-agent-root/bamboo-agent-home/temp/518946826-518455357-517442402-ScriptBuildTask-7080740712063035315.sh: 13: /opt/bamboo-agent-root/bamboo-agent-home/temp/518946826-518455357-517442402-ScriptBuildTask-7080740712063035315.sh: oc: not found
26-Jul-2019 20:21:27    /opt/bamboo-agent-root/bamboo-agent-home/temp/518946826-518455357-517442402-ScriptBuildTask-7080740712063035315.sh: 13: /opt/bamboo-agent-root/bamboo-agent-home/temp/518946826-518455357-517442402-ScriptBuildTask-7080740712063035315.sh: oc: not found

，这就是我要执行的:-

PATH=$PATH:/data/bamboo/openshift
echo "-> Getting present working dir..."
echo $PWD
oc project dummyproject-tst
echo "-> Deployment started..."
echo "-> Bringing down the pod(s)..."
oc scale --replicas=0 deployment dummy-deployment
echo "-> Recreating the pod(s)..."
oc process -f ./openshift/templates/dummy-template.yaml -p IMAGE_TAG=190724.262 | oc apply -f -
echo "-> Deployment completed on TST environment"

我想，openshift cli的环境变量不起作用，尽管我已经在第1行声明了PATH=$PATH:/data/竹/openshift

Answer 1

当我遇到这样的问题时，通常是因为竹类最终会有一个不登录的shell，所以您的正常PATH永远不会构建。(有关什么是非登录shell的更多信息，请参见https://unix.stackexchange.com/a/46856/36922)。我用了几种方法来解决这个问题：

1. 使用source命令在脚本开头运行/etc/profile。(YMMV在实际路径上--这在Ubuntu上工作)。

source /etc/profile

1. 在启动命令时使用-l标志手动强制登录shell。最简单的方法是将实际脚本转储到文件中，然后手动执行。就像这样：

echo "echo Hi!" > /tmp/tmp.sh
/bin/sh -l /tmp/tmp.sh

我推荐选项1而不是选项2--当其他人阅读您的脚本并有问题时，它会更易于维护，也不会让人感到困惑。