I have a powershell script that runs for ~30 minutes (waiting on various process' to finish). At the end, it writes a message to the event log, determining if the process was a success or failure. I plan on hosting this script on teamcity and want the build to fail, but don't know how to handle the interaction between the script and teamcity in order for this to happen.
I'm looking for a way to make a powershell script that is ran remotely to communicate to teamcity whether it was a failure or success. I've read up on a lot of the teamcity documentation and I'm still not sure how to start going about this.
You should probably consider using TeamCity Service Messages, or specifically Reporting Build Problems.
An example of how to emit a service message using PowerShell (assuming you're using the PowerShell build step):
Write-Output "##teamcity[buildStatus text='I am a successful build']"
or
Write-Output "##teamcity[buildProblem description='$powershell_error_message']"
where you can inject the captured powershell error message.
To intentionally fail a build in teamcity you should use "[Environment] :: Exit (1)" instead of "Exit 1".
More information can be obtained at the link below:
https://confluence.jetbrains.com/display/TCD9/PowerShell
Related
Rundeck 4.8.0 community version on Redhat 9 Linux with Windows node.
My Rundeck jobs call powershell (.ps1) scripts on the windows node.
If there are any errors encountered in the script, the Rundeck job dies.
The rundeck output gives the NonZeroResultCode message
NonZeroResultCode: [WinRMPython] Result code: 1
There's more code that needs to be run after where the error occurred, but Rundeck just dies and doesn't continue the rest of the .ps1.
I previously used Rundeck version 3.something, I thing it was 3.9.
If there was an error in the script, such as a get or a set failed, the Rundeck console would just display the text of the error in red, and continue.
Now I know I can change my code and add try/catch statements, -erroraction SilentlyContinue and so on. However it makes no sense to me that Rundeck takes it upon itself to kill my script because a get or a set failed.
I want to be the one to decide if I want to exit the script or not, I don't want Rundeck to make that decision.
Can this behavior be changed?
thanks in advance.
That's the default Rundeck behavior.
You can "attach" an error handler on that script-step (on any step actually), e.g: the error handler could be the script code when your step fails.
The error handler feature is designed for that kind of scenario, take a look at this.
I'm new to the whole Azure DevOps world and just got transferred to a new team that does just that.
One of my assignments is to fix an issue with a pipeline where one of the steps runs a shell script that installs an application. Currently, the step seems to run without any issue shown on the log, but when we connect to the container's pod, the app is not there.
If we run the script directly inside the pod, the application is installed correctly. I'm not sure how to track this. One of the things I've tried was to check the event log to see if there's any error while the installation is executed:Get-Eventlog -LogNmae "Windows PowerShell" -Newest 20, so far no luck here. Again, kinda of new at this, not sure what other tools are out there to track the reason why the script is not installing during the pipeline execution.
To troubleshoot your pipeline run, you can configure your pipeline logs to be more verbose.
1, To configure verbose logs for a single run, you can start a new build by choosing Run pipeline and selecting Enable system diagnostics, Run.
2,To configure verbose logs for all runs, you can add a variable named system.debug and set its value to true.
You can also try logging into your agent server and check for the event log. See this blog for view event log on windows.
The issue was related to how the task was awaited. Adding this piped params helped us solve the issue:
RUN powershell C:\dev\myprocess.ps1 -PassThru | Wait-Process;
I am trying to have a status like failed or pass to be passed to Octopus to fail or pass a deployment.
I have three different projects, one project per product. Each of these tests are run today on the server using batch scripts. I am trying to work on a solution using powershell script, which will invoke these tests as batch and if the test results fail pass a variable and value back to Octopus to fail our deployment. We do not want our code to deploy in stage if tests fail.
I have bunch of scripts I am trying, like get Exitcode, but nothing seems to be working. I have the Scripts working if I kick off the .ps1 on the server. But I need the ability to pass fail or pass status to Octopus. can someone please help?
I'm looking for the best solution for automating Windows server 2012r2 via Rundeck. I'm extremely familiar with Rundeck but use it for linux vms. I've searched online for this topic but havent found anything that seems reliable. Ideally, I want to start a Runeck Job that can add users in Active Directory but not sure how to approach it. I've tried using winrm but have gotten mixed results. Ironically, it will run basic powershell commands but it errors when trying to do anything with AD. Ive even tried creating a powershell script on the AD server and have Rundeck simply execute the powershell script to no avail. Hopefully someone has had success in controlling windows nodes with Rundeck. Below is the error i receive when trying to run a powershell script.
Execution failed: 27 in project windows: [Workflow result: , step failures: {1=Dispatch failed on 1 nodes: [Windows_AD_Server: NonZeroResultCode: [WinRMPython] Result code: 1 + {dataContext=MultiDataContextImpl(map={ContextView(node:Windows_AD_Server)=BaseDataContext{{exec={exitCode=1}}}, ContextView(step:1, node:Windows_AD_Server)=BaseDataContext{{exec={exitCode=1}}}}, base=null)} ]}, Node failures: {Windows_AD_Server=[NonZeroResultCode: [WinRMPython] Result code: 1 + {dataContext=MultiDataContextImpl(map={ContextView(node:Windows_AD_Server)=BaseDataContext{{exec={exitCode=1}}}, ContextView(step:1, node:Windows_AD_Server)=BaseDataContext{{exec={exitCode=1}}}}, base=null)} ]}, status: failed]
Firstly, PowerShell already offers a built-in way to execute jobs, with no 3rd party addons.
About Jobs
Provides information about how PowerShell background jobs run a command or expression in the background without interacting with the current session.
About Remote Jobs
Describes how to run background jobs on remote computers.
PowerShell Jobs Week: Remote Jobs
I've never heard of / used Rundeck as Paul points out as well, so this just be the Rundeck has particulars that need to be in play first. Yet, looking at the docs, and a quick youtube video on the topic, there are several things that must be in place for what you say here...
I've tried using winrm but have gotten mixed results.
… to work.
Video - Running commands remotely using the console with WinRM/WinRS
If Rundeck is similar to SCCM or Scheduled Task, then the same approach applies.
Write Your PowerShell script. have Rundeck/ScheduedTask call powershell.exe to run the script
How to execute a PowerShell script automatically using Windows task scheduler?
Also, this could very well be seen as a duplicate of this stackoverflow discussion and answer.
rundeck unable to execute powershell script with import-module
I am setting up somethin similar and have the same problem. The AD operation executes successfully, but returns exit code 1 and throw the error you mentioned. I am in the process of debugging the plugin, but it seems there is a bug within response handling.
Please check if the ad operation works although the error is thrown and please post the script block you are using to control the AD.
Edit: in meantime I was able to narrow down the issue and opened an issue with the project: github
I provisionally fixed the issue with adding the following line in winrm-session.py after line 89:
new_msg = msg
The line before is:
" error message: %s" % (e))
The line after is:
else:
If this does not solve your issue, please post your script block.
Best Tobias
I've developed a Powershell script to deploy updates to a suite of applications; including SQL Server database updates.
Next I need a way to execute these scripts on 100+ servers; without manually connecting to each server. "Powershell v2 with remoting" is not an option as it is still in CTP.
Powershell v1 with WinRM looks the most promising, but I can't get feedback from my scripts. The scripts execute, but I need to know about exceptions. The scripts create a log file, is there a way to send the contents of the log file back to the "client" (the local computer making the remote calls)?
Quick answer is No. Long version is, possible but will involve lots of hacks. I developed very similar deployment script/system using PowerShell 2 last year. The remoting feature is the primary reason we put up with the CTP status. PowerShell 1 with WinRM is flaky at best and as you said, no real feedback apart from ok or failed.
Alternative that I considered included using PsExec, which is very much non-standard and may be blocked by firewall. The other approach involves using system management tools such as MS's System Center, but that's just a big hammer for a tiny nail. So you have to pick your poison...
Just a comment on this: The easiest way to capture powershell output is to use the start-transcript cmdlet to pipe console output to a file. We have a small snippet at the start of all our script that sends a log file with the console output from each script to a central file share, and names the log file with script name and date executed so that we'll have an idea of what happened. Its not too hard to pipe all those log files into a database for further processing either. Probably won't seolve all your problems, but would definitely help on the "getting data back" part.
best regards,
Trond