Fix flaky test: TestBrutalShutdown.testQueryRetryOnShutdown

[ZacBlanco]

Description

Set error code REMOTE_BUFFER_CLOSE_FAILED as a retriable error.

(Hopefully) Fixes #22125

Motivation and Context

I set this test's invocationCount to 10 or 15 in order to run this on repeat and trigger the failure. I found that for all failures, it was because the query failed with REMOTE_BUFFER_CLOSE_FAILED. This caused the test to fail because it is not marked as a retriable error. By changing the respective error code to one that is retriable this test no longer seems to fail. I was not able to reproduce the failures on my local machine

This may have occurred due to changes in the way the task event loops are scheduled, but I am not 100% sure. Either way, I think this error code is capable of query retries, so the solution seemed suitable.

Impact

• StandardErrorCode#REMOTE_BUFFER_CLOSE_FAILED is now a retriable query error.

Test Plan

Existing tests

Contributor checklist

• Please make sure your submission complies with our contributing guide, in particular code style and commit standards.
• PR description addresses the issue accurately and concisely. If the change is non-trivial, a GitHub Issue is referenced.
• Documented new properties (with its default value), SQL syntax, functions, or other functionality.
• If release notes are required, they follow the release notes guidelines.
• Adequate tests were added if applicable.
• CI passed.

Release Notes

== RELEASE NOTES ==

SPI Changes
* REMOTE_BUFFER_CLOSE_FAILED is now a retriable error

[steveburnett]

Nit suggestion for the release note entry to follow the Order of changes in the Release Notes Guidelines:

== RELEASE NOTES ==

SPI Changes
* Improve testing by making REMOTE_BUFFER_CLOSE_FAILED a retriable error

I think my "testing" is a little vague. If you have a better phrasing, please do!

[jaystarshot]

Don't think we should add REMOTE_BUFFER_CLOSE_FAILED as an retryble error just for unit tests. We can add it if its indeed needed in prod too.
cc: @rschlussel

[jaystarshot]

That being said i don't see many errors REMOTE_BUFFER_CLOSE_FAILED internally so its probably fine

[ZacBlanco]

Another solution is to try and re-write the test assertion to assert that all queries which initially failed with retriable errors all eventually successfully pass - but it's possible you can get 0 queries with retriable errors, so it would defeat the purpose of the test in some runs. The query failure type is non-deterministic when the worker(s) shut down, which seems to be the main cause of flakiness.

I am open to other ideas

[rschlussel]

I think if this is an error code that you see after a forced shutdown, it makes sense to mark it as retriable.