ludown : Sorry unable to open [general.lu]

[bfbatch]

Project

Name: Virtual Assistant-->

Language: C#

Description

While running the deploy_cognitive_models script, it errors unable to find the .lu files

To Reproduce

Either run deploy.ps1 or deploy_cognitive_models.ps1

Expected behavior

Finds file and deploys app

Screenshots

Additional context

Looks like maybe a path problem. Not sure. Here is what I'm running

.\Deployment\Scripts\deploy.ps1 -name xxxpoccbot -resourcegroup n21-xxx-poc-cbot-RGRP -subscription xxx -location centralus -luisauthoringkey xxx -luisauthoringregion westus -parametersFile .\Deployment\Resources\parameters.xxx.json -appsettingsfile .\appsettings.xxx.json -PSScriptRoot C:\XXX\XXX.Chatbot\XXX.Chatbot\Deployment -languages en-us
-projDir .\ -appId XXX -appPassword XXX

[stevkan]

Can you confirm for me the direction in which you are running the command? Also, can you verify that they file exists?

[bfbatch]

Hi Steven, I’m running the command from PS ISE locally against the repo root. Yes they exist in the Deployment/Resources folder. Get Outlook for iOS On Fri, Jun 28, 2019 at 9:15 PM -0400, "Steven Kanberg" <notifications@github.com> wrote: Can you confirm for me the direction in which you are running the command? Also, can you verify that they file exists? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

[bfbatch]

Hi Steven, Thanks for the reply. I'm running the script from PS ISE and using the deploy.psi in the scripts folder to deploy. Yes the .lu files are there as I've cloned the VA template.  Regards,Brian Batchelderc-850-228-5462 On Friday, June 28, 2019, 09:15:50 PM EDT, Steven Kanberg <notifications@github.com> wrote: Can you confirm for me the direction in which you are running the command? Also, can you verify that they file exists? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

[darrenj]

One quick question which might help @stevkan and I - Does your full path have any spaces in it (e.g. space in a user name or similar?)

[bfbatch]

Hi Darren, neither the name nor the path to the files have spaces.

[stevkan]

@bfbatch, sorry for the delay. Can you also tell me which version of PS you are running this on?

[bfbatch]

@stevkan np. I'm using 5.1.1. since the client wanted this to run in Azure DevOps and that is the version running there.

[Zerryth]

Just followed the Virtual Assistant readme as it is today, starting from scratch, and was able to get the deploy.ps1 command to run successfully. -- wasn't able to repro the behavior that @bfbatch is experiencing

I basically installed all the requirements up-to-date, used the VSIX template in Visual Studio to create the VA bot template, navigated down and ran this command, looks like 3 VA folders in:

---

@bfbatch, could you try installing PowerShell Core version 6, as specified in the readme, and see if that helps resolves the issue?

Also to be sure to update the other requirements as well (such as AZ CLI, and the other CLIs specified in the readme)

[CoHealer]

Hi @bfbatch, have you had a chance to try the recommendation @Zerryth put forth?

[bfbatch]

@CoHealer @Zerryth Thanks for the reply. I have not had a chance to test again, but will tonight and let you know.

[threeyay]

Hi, I am getting the same error, my user name has a space in it and hence the path has a space in it too.
When creating the project with the template, it does not create the Deployment\Resources\LU\en\general.luis,
Deployment\Resources\QnA\en\chitchat.qna, and
Deployment\Resources\QnA\en\faq.qna,
files.
The deploy_cognitive_models_log.txt file says these files are not found but they were not created in the first place.

[Zerryth]

@threeyay please follow the other thread you opened with regards to the issue (#1923) you are experiencing, as the reason for not deploying correctly would be different from what bfbatch is experiencing since he has the deployment resources.

This way we can separate any confusion that may arise between the resolutions to different problems

[Zerryth]

For reference, these are the versions of the dependencies that I'm using, referenced in the readme that specify that you have updated versions:

Visual Studio 2017
Virtual Assistant Template v 4.4.4
.NET Core 2.2
npm 6.10.0
PowerShell Core Version 6
(bot) dispatch 1.5.1
ludown 1.3.2
luis-apis 2.6.0
qnamaker cli 1.3.1
luisgen 2.2.0.0
az cli 2.0.69 (minimum required is 2.0.64!)
Bot Framework Emulator v4.5

@bfbatch let us know what you find when you try again

[bfbatch]

Hi @Zerryth I started a new VM and installed/updated to the above dependencies and cloned the 4.4.4 VA template. Currently it hangs at "Validating Azure Deployment"

[Zerryth]

Hmm...ok. Let's try these things:

1. Upgrade to Virtual Assistant VSIX template 4.5.1 (just released)
2. Instead of providing all the parameters in one command as you did (which should in theory work), let's instead just try following the prompts to provide the parameters
3. If deployment still fails, take a look at the deployment_log.txt file and let us know if you're getting any errors there

(Other potential issues could be perhaps in "." in path, special characters in appPassword, or your override on PSScriptRoot)

[bfbatch]

Still no go. Same results for all of the above. I also generated a new key for the app registration to remove the '.', but still hangs at "Validating Azure Deployment". + it doesn't get far enough to create the deploy_log.txt. Anything else I can try?

[Zerryth]

Going to circle back with the solutions team, and I'll let you know what else we could try

[Zerryth]

@bfbatch, still waiting to hear back from the team, however before I connect with them, wanted to verify: you mentioned switching to a VM, and I noticed that in your original screen shot that you provided, it actually got passed the "Validating Azure Deployment." prompt

Did you make sure to log into your account in az cli before running the deployment script? (Quick way to check is az ccount show, which should output something like this:

{
  "environmentName": "AzureCloud",
  "id": "xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "isDefault": true,
  "name": "Bot Framework Support Team",
  "state": "Enabled",
  "tenantId": "xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx",
  "user": {
    "name": "xxxxxxxxxxxxx@microsoft.com",
    "type": "user"
  }
}

If you're not logged into az cli, you can do so with the az login command, then run the deployment script.

---

Separately, and I don't think this is the issue, but for the sake of trying, make sure that when you're running the deployment script, you're using a unique name, as sometimes resources might have been partially created in the resource group. You can delete the resources created and "wait a while" and reuse the same name, however I do know it does take a little bit for the resources to delete

[jwiley84]

I'm helping @Zerryth look into the issue, but I do want to emphasize something she pointed out:

Separately, and I don't think this is the issue, but for the sake of trying, make sure that when you're running the deployment script, you're using a unique name, as sometimes resources might have been partially created in the resource group. You can delete the resources created and "wait a while" and reuse the same name, however I do know it does take a little bit for the resources to delete

Please make sure you're taking the time after every failed deployment to delete anything that was partially created. You will have empty resource groups and partially created resources for every run of the script that WILL hiccup further attempts to redeploy, especially using the same names.

[CoHealer]

@bfbatch, have you confirmed that you are scrubbing your resources after the failed deployments? Are you still getting a repro?

[bfbatch]

Hi Jessica, Thanks for the assist. Yes, I was deleting the previous resource groups. Here is the history of the scripts. 1. Initial attempt was to run them from Azure DevOps. The only PS tasks I could find was for vs 5.1.1. Once I change the required version from within the scripts I was able to deploy the environment. However, I was getting the error that the .lu files were not found. 2. After submitting a issue, I was advised to make sure I was following these requirements. So I created a new VM and started from scratch. I installed all the below and then cloned the repo again. When I run the scripts now, it hangs at Validating. I have tried to initiate an error by removing the appsettings file and it did error. Problem is there is no error or deploy.txt errors, nothing to indicate what is happening. Is there anyway to get a debug or verbose during this process? Or a screen sher to make sure I'm not doing something wrong? I'g going to attempt to step through the code, but I'm not sure it will help if I can't get any process logs. Please advise. - Visual Studio 2017 - Virtual Assistant Template v 4.4.4 - .NET Core 2.2 - npm 6.10.0 - PowerShell Core Version 6 - (bot) dispatch 1.5.1 - ludown 1.3.2 - luis-apis 2.6.0 - qnamaker cli 1.3.1 - luisgen 2.2.0.0 - az cli 2.0.69 (minimum required is 2.0.64!) - Bot Framework Emulator v4.5 Regards,Brian Batchelderc-850-228-5462 On Tuesday, August 6, 2019, 11:26:19 AM EDT, Jessica Wailes <notifications@github.com> wrote: I'm helping @Zerryth look into the issue, but I do want to emphasize something she pointed out: Separately, and I don't think this is the issue, but for the sake of trying, make sure that when you're running the deployment script, you're using a unique name, as sometimes resources might have been partially created in the resource group. You can delete the resources created and "wait a while" and reuse the same name, however I do know it does take a little bit for the resources to delete Please make sure you're taking the time after every failed deployment to delete anything that was partially created. You will have empty resource groups and partially created resources for every run of the script that WILL hiccup further attempts to redeploy, especially using the same names. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[jwiley84]

Hi Brian!

I know it's annoying to get the constant "can you confirm xyz", but...

can you confirm that the deployment script IS or IS NOT creating a deploy_log.txt file? When running the script as outline in the steps (simply navigating to your project directory and running .\Deployment\Scripts\deploy.ps1) the file SHOULD be created in your Scripts folder. It would be initially empty.

[bfbatch]

Hi Jessica,

Np, I completely understand. The deploy.txt file was created when I generated a real error once I removed the appsetting file. Prior to that, the deploy.txt file didn't get created when the script hangs at "validating deployment".

[bfbatch]

Hi Jessica,

I've got a new update. As of just now I was able to deploy using the VA template. :-) It looks like the validation was hanging on something in the base template that the script didn't like. When I deployed using the original base template (default from VA unmodified) the environment and cognitive models deployed. My updated base template does pass JSON validation and I re-deployed the same template to the environment using a standard AzureRM deployment script and it was successful. Can someone review my template to see if they can spot what the deploy script was hanging on?

[jwiley84]

Brian,

I apologize, I've been out of the office for the last few days. Do you mean the template.json?

[bfbatch]

hi Jessica,
Yes I'm referring to the template.json.

[jwiley84]

Go ahead and post it, let's see what's going on.

[bfbatch]

Thanks Jessica
template.json.zip

[jwiley84]

Thank you for your file. I ran it through a diff checker to see what kinds of changes (and ignore the unchanged parts), and there are 75+ changes to that file. At a glance, I can't tell which single change would have hung up the validator, and I recommend you take the generic template (which you said above works fine) and add in your new changes you made, one at a time, testing them against az group deployment validate to see the results.

[bfbatch]

Thanks Jessica,
Yes I knew there were about 100 changes, and I run these type of templates through Azure deploy scripts all the time. Wondering what is the main difference between the validator in the chatbot depoy and Azure deploy (besides AzuerRm vs Az). This was for a client and as I mentioned I just run the same template via AzureRm deploy to deploy all the changes and we are up and running. Since I don't have access to the clients subscription, I'll need to run my validation checks in my own. Would you like to close this issue or leave it open for my feedback once I run my checks?

[darrenj]

Thanks for the work on this.

We're just using the standard az cli capabilities to validate (az group deployment validate so there should be no difference between the two. You might want to verify the az versions using az-v and ensure they are >=2.0.64 too.

I just ran this az group deployment validate --resource-group djrg1 --template-file template.json and given I don't have your parameter file provided a bunch of manual settings and it was able to validate.

When you call the deploy script I see you are passing a parameters file - could you provide an example of this or try calling the same CLI command we do directly:

az group deployment validate `
		--resource-group $resourcegroup `
		--template-file "$(Join-Path $PSScriptRoot '..' 'Resources' 'template.json')" `
		--parameters "@$($parametersFile)" `
		--parameters name=$name microsoftAppId=$appId microsoftAppPassword="`"$($appPassword)`"" `
        --output json

[bfbatch]

Thanks @darrenj I will run through the validation and report back.

[CoHealer]

@bfbatch, is there any update?

[bfbatch]

@CoHealer I've validated my parameters file, and I'm stepping through my edited template to see what is causing the validation to hang. Should have a final report later today.

[lauren-mills]

@bfbatch - Wanted to check back in and see if you had any more details. If not, we'll close this due to inactivity.

[bfbatch]

@lauren-mills Thanks for the push. I did walk through my custom template, on resource resource at a time. I didn't hit any errors or hanging of the validation on individual resources, only when the entire template was used. We can close this as I'll be starting a new client to deploy a new environment and I'll follow up with any updates.

[lauren-mills]

Thanks, @bfbatch! Please let us know.