I would like to create an image containing nanoiis image with the sqlcmd utility installed. I could further use this custom image in our DevOps pipeline.
The sqlcmd is available at [this sqlcmd link] from the microsoft website
https://learn.microsoft.com/en-us/sql/tools/sqlcmd-utility?view=sql-server-2017.
I'm kind of stuck because the utility is in a msi file (MsSqlCmdLnUtils.msi) and I don't know how to run a software that would popup an UI unlike SQLserver that can be installed from command line.
Once done, I can commit the image and use it in our pipeline.
2 things:
Nano server does not support IIS role anymore
You can not install MSI files on nano server either
You need to base your image off windowsservercore instead and run your installation in quiet mode to avoid popups in MSI file.
Related
List item
pipelines:
default:
- step:
name: Push changes to Commerce Cloud
script:
- dcu --putAll $OCCS_CODE_LOCATION --node $OCCS_ADMIN_URL --applicationKey $OCCS_APPLICATION_KEY
- step:
name: Publish changes Live Storefront
image: Python 3.5.1
script:
python publishDCUAuthoredChanges.py -u $OCCS_ADMIN_URL -k $OCCS_APPLICATION_KEY
environment variables:
$OCCS_CODE_LOCATION: Path to location of all OCCS code
$OCCS_ADMIN_URL: URL for the administration interface on the target Commerce Cloud instance
$OCCS_APPLICATION_KEY: application key to use to log into the target Commerce Cloud administration interface
So I want to use Azure Dev Repository to CI / CD.
in the above code block if you see I have specified - dcu & python code in two task.
dcu is nodejs third party oracle tool which needed to be used to migrate code to cloud system. I want to know how to use that tool in azure dev ops,
Second python (or) nodejs which I want to invoke to REST api to publish the changes.
So where to place those files and how do we invoke it.
*********** Update **************
I hosted the self pool agent and able to access the system.
Just start executing basic bash code, but end up in two issue -
1) the git extract files from the repository it is going to _work/1/s, not sure how that path is decided. How can I change that location s
2) I did 'pwd' to the correct path but it fails in 'dcu' command. I tried with npm and other few commands it fails. But things like mkdir , rmdir it create & remove folder correctly from the desired path. when I tried the 'dcu' cmd from the terminal manually from the system it works fine as expected.
You can follow below steps to use DCU tool and python in azure pipelines.
1, create a azure git repo to include dcu zip file and your .py files. You can follow the steps in this thread to create a azure git repo and push local files to azure repo.
2, create azure build pipeline. Please check here to create a yaml pipeline. Here is a good tutorial for you to get started.
To create a classic UI pipeline, please choose Use the classic editor in the pipeline setup wizard, and choose start with an Empty job to start with an empty pipeline and add your own steps.(I will use classic UI pipeline in below example.)
3, Click "+" and search for Extract files task to unzip the DCU zip file. Click the 3dots on the Destination folder field to select a destination folder for extracted dcu files. eg. $(agent.builddirectory). Please check my answer in this thread more information about predefined variables
4, click "+" to add a powershell task. Run below script in screenshot to install dcu and run dcu command. For environment variables (like $OCCS_CODE_LOCATION), please click the variables tab in below screenshot to define them
cd $(agent.builddirectory) #the folder where the unzipped dcu files reside. eg. $(agent.builddirectory)
npm install -g
.\dcu.cmd --putAll $(OCCS_CODE_LOCATION) --node $(OCCS_ADMIN_URL) --applicationKey $(OCCS_APPLICATION_KEY)
5, add Use python version task to define a python version to execute your .py file.
6, add Python script task to run your .py file. Click the 3dots on Script path field to locate your publishDCUAuthoredChanges.py file(this py file and the dcu zip file have been pushed to azure git repo in the above step 1).
You should be able to run the script of above question in the azure devops pipeline.
Update:
_work/1/s is the default working folder for the agent. You cannot change it. Though there are ways to change the location where the source code is cloned from git, the tasks' workingdirectory is still from the default folder.
However, You can change the workingdirectory inside the tasks. And there are predefined variables you can use to refer to the places in the agents. For below example:
$(Agent.BuildDirectory) is mapped to c:\agent_work\1
%(Build.ArtifactStagingDirectory) is mapped to c:\agent_work\1\a
$(Build.BinariesDirectory) is mapped to c:\agent_work\1\b
$(Build.SourcesDirectory) is mapped to c:\agent_work\1\s
The .sh scripts in the _temp folder are generated automatically by the agent which contains the scripts in the bash task.
For above dcu command not found error. You can try adding dcu command path to the system variables path for your local machine's environment variables. (path in user variables cannot be found by agent jobs, For the agent use a different user account to connect to local machine)
.
Or you can use the physically path to dcu command in the bash task. For example let's say the dcu.cmd in the c:\dcu\dcu.cmd on local machine. Then in the bash task use below script to run dcu command.
c:/dcu/dcu.cmd --putAll ...
I am trying to debug an application I have deployed to google cloud app engine. Reading the docs, I figured out that in order to do so I have to enter the debug mode using
gcloud app --project [Project ID] instances enable-debug
afterwards I am able to SSH into my instance and access root. Now I would like to edit some of the files. However, trying to use vim or nano does not seem to work.
Is there a way to edit those files without re-deploying the entire app?
Once you SSH into the App Engine instance and open a shell into the Docker container, you'll need to download the package list before installing nano or vim:
apt-get update && apt-get install nano
Then you can edit your app's files (which are in /app):
nano composer.json
The deployed app runs live code. It is not generally feasible to edit it. Moreover, changes made to the running container are not permanent; in fact they and are lost at the first re-start.
You may find some information on the Debugging an Instance page.
Unrelated to the above, an actual command-line editor is offered in the cloud shell.
I have a custom Java app and an IzPack installer. For years, in my izpack build file I had the following:
<run-privileged condition="izpack.windowsinstall.vista|izpack.windowsinstall.7"/>
The problem is that some of the users do not have admin privilege on their PCs, but they still want to be able to install the package. If I remove the above, they can run the installer but then it complains "This directory cannot be written!", when they try to install in the default location, which is C:\OPENDCS.
Yet the same user can create this directory either from a CMD or an Explorer window.
Is there a way to allow the izpack installer to create a directory directly under C:\ without running as an administrator?
Please check the behavior with izpack v5.0.7. The problem you mentioned should be fixed with this issue: https://izpack.atlassian.net/browse/IZPACK-1355
You could package your directory create operations in a create-dirs.bat batch file, which you would mark <executable> and execute stage="postinstall". This way the directory creation will be executed with the given user's permissions, which (according to your post) should work just fine.
EDIT 29/02/2016: You would put this file into a first "dummy" <pack>, mark it <executable> and execute stage="postinstall" as stated above, which would execute it after this first dummy pack was installed. At the installation of the next pack (i.e. your first useful pack) you will already have the folder created.
Note that postinstall will not run the batch file after the installation, but after the <pack>'s installation.
I am currently working on a Windows batch file that will allow me to silently install git (the executable for which will be placed in the folder that the .bat file will be running from) in a pre-specified location on the file system.
I've found this article which seems to provide some suitable advice:
https://github.com/msysgit/msysgit/wiki/Silent-or-Unattended-Installation
However, I'm not entirely sure what parameters I would need to mention in my LOADINF file. I would like to pre-define the options that the user would manually select throughout the various stages of installation, so that it can run through from start to finish without prompting anything from the user.
Can anyone help or point me to a place where I can find these parameters and their available values?
Create a file, for eg. my-config.cnf (or my-config.ini) with the following content:
[Setup]
Lang=default
Dir=C:\Program Files (x86)\Git
Group=Git
NoIcons=0
SetupType=default
...
<other options as shown in the msysgit wiki>
Now, in the batch file, when you execute the installation file (say msysgit-install.exe), use /LOADINF as follows:
msysgit-install.exe /SILENT /LOADINF="my-config.cnf"
I have an old setup of the old program written on c++ which contains multiple installation files files.
_SETUP.1
_SETUP.DLL
_INST32I.EX_
_ISDEL.EXE
SETUP.EXE
DISK1.ID
SETUP.INI
SETUP.INS
_SETUP.LIB
SETUP.PKG
I want to combine all that in to single executable file and i want to execute SETUP.EXE when user would run that single executable. Is it possible to achieve somehow?
The easiest way is simple create archive and say to user to to unpack that and to run SETUP.EXE but i am just wondering may be i can create setup like i describe above.
IExpress.exe is ideal for your job. Google for samples. It is included in your Windows installation. Just open a Command Prompt and type iexpress.exe - this starts a wizard that helps you getting started.