Metadata API odd errors during the long pulling

[steals]

Summary

This issue is the continuation of #1835

During our daily deployment, we have a lot of failures because in the middle of the pulling results Metadata API hangs with some odd errors, like:
ERROR running force:source:deploy:report: Metadata API request failed: request to https://.....sandbox.my.salesforce.com/services/Soap/m/56.0 failed, reason: Client network socket disconnected before secure TLS connection was established
or
ERROR running force:source:deploy:report: Metadata API request failed: Unexpected internal servlet state
or
ERROR running force:source:deploy:report: Metadata API request failed: Unable to complete the creation of the query cursor at this time, please retry your query later.

Is it possible to add these errors to the list of ignored errors so the sfdx cli doesn't stop pulling for the results?

System Information

sfdx-cli/7.182.1 linux-x64 node-v18.12.1

running on Ubuntu 20.04 docker image

[github-actions]

Thank you for filing this issue. We appreciate your feedback and will review the issue as soon as possible. Remember, however, that GitHub isn't a mechanism for receiving support under any agreement or SLA. If you require immediate assistance, contact Salesforce Customer Support.

[WillieRuemmele]

I haven't seen those errors before, we can add them to the retry-able list

[git2gus]

This issue has been linked to a new work item: W-12408010

[steals]

@WillieRuemmele any ETA when this issue can be resolved?

[mshanemc]

@steals this'll be in RC on Thursday Feb16 and the main channel the following week (Feb23).

[gfarb]

Commenting to subscribe but wanted to call out at GitHub we have been running into these errors a lot recently, primarily:

• Metadata API request failed: Unexpected internal servlet state
• Metadata API request failed: Failed while fetching query cursor data for this QueryLocator

It causes our GitHub Actions to fail and despite the fact that the deployment continues, we need to re-run our CI/CD jobs which results in time and resources wasted

[hudec117]

@mshanemc that PR doesn't seem to address this error message?

Metadata API request failed: Failed while fetching query cursor data for this QueryLocator

This is the main one we're seeing at the moment.

[steals]

Also noticed this new error today. I guess we need to create a new GitHub ticket for that.

[mshanemc]

that last one will go out in tomorrow's RC.

[shetzel]

Here is some additional information about the server side errors you're seeing during polling for deploy status:

Failed while fetching query cursor data for this QueryLocator - There is a known bug tracked by the appropriate internal team. W-12563567. Feel free to open support cases and mention that to bump the priority for that team.

Unexpected internal servlet state - This looks like a known issue: https://issues.salesforce.com/issue/a028c00000gAzYKAA0/intermittent-unexpected-internal-servlet-state-during-soap-and-bulk-api-requests

Happy to add them to the list of "retry-errors". Letting support know about them will also help prevent them from happening at all.

Thanks!

[hudec117]

We're using sfdx-cli/7.190.2 linux-x64 node-v18.14.1 in our CI/CD pipelines and we're still seeing the Metadata API request failed: Failed while fetching query cursor data for this QueryLocator error that ends the SFDX process and fails the build.

Was this fix included in 7.190.2 or in the upcoming 7.191.1?

[cristiand391]

@hudec117 sorry for the late reply, v7.190.2 should definitely include this fix. Can you open a new GH issue with the commands you are running? Thanks!
https://github.com/forcedotcom/cli/issues/new/choose