Attaching an artifact freezes the deployment

Hi,

We’ve recently build a new test environment which we are deploying our DACPACs to. Part of our deployment process is to attach deploy reports and deploy scripts to the release so that we can see changes coming from sqlpackage.exe before we approve the actual deployment.

We have this process working correctly in our UAT and Production environments but when we try to run the step that attaches the artifacts it stalls and eventually get a timeout error that the tentacle cannot communicate on port 10943.

Our tentacles are polling tentacles connected via a proxy and our deploys work fine if we don’t attach artifacts.

Is there some proxy config we need to do or is this a permission issue on the tentacle box?

Hi John,

Thanks for getting in touch.

Am I right in understanding that attaching artifacts is working correctly in your UAT and Production environment but is failing in the new test environment?

If you are attaching large files that take a long time to transfer, the transfer may be timing out. Would you mind sending the raw task log for a deployment where this is happening? (https://octopus.com/docs/how-to/get-the-raw-output-from-a-task)

Cheers,
Shane

Hi Shane,

Sorry for the delay here.

Yep, attaching artifacts after we run sqlpackage.exe works fine in UAT and Prod but this new test environment fails.

These files are less than 1MB; they are the deploy report xml and sql script.

I’ve attached the raw output from the failed step.

octo-artifact-fail.txt (5 KB)

Hi John,

Thanks for the reply, it makes a lot more sense now.

You are hitting our 10 minute timeout when processing the polling request. Usually this happens for large files and we can just bump the timeout up but I don’t think it will help in this case, as you alluded to in the original post it is probably a network issue.

You could try increasing the logging level on one of your Tentacle and sending the logs. There are instructions here: https://octopus.com/docs/reference/log-files#Logfiles-Changingloglevels
Instead of octopus.server.exe.nlog you would need to change tentacle.exe.nlog located by default in C:\Program Files\Octopus Deploy\Tentacle.

I would also recommend capturing the traffic on the Tentacle machine using a utility like WireShark (https://www.wireshark.org/) to see if you can spot what is happening to the artifact upload.

Cheers,
Shane

Hi Shane,

Thanks for info; we will enable more logging and take a look.

Is there anyway to adjust the polling tentacle interval for checking polling requests? Keen to try more than 10 seconds to test a few things out.

Hi Shane,

We sorted this out finally. It was due to a security group not allowing anything other than TCP connections.

Our tentacles in the new environment are fine now.

Hi John,

Great to hear, thanks for letting me know.

Cheers,
Shane