Redis startup race condition - Not setting up DB credentials

[jpwhitemn]

🐞 Bug Report

Affected Services

Continuation from #2542 - while fix worked for Geneva, there was a suggestion on the part of Tony and Andre needed for Hanoi going forward:

Tony: Re: [ -s ], isn't it still possible for the the file to be partially written when this check is made? Wouldn't a semaphore file be safer?
Andre: Given the length of the password and the typical I/O buffer, it is unlikely for a partial write to occur.

When Geneva is released, I'm going to rewrite security-secretstore-read to eliminate the password file altogether and set the password directly in Redis.

Is this a regression?

temp solution in Geneva

Description and Minimal Reproduction

see #2542

[hutchic]

decided this is a duplicate of #2636 going to track progress on that issue going forward

edit: I meant to link to #2503

[andresrinivasan]

@hutchic Did you mean to close this issue (#2636) or a different one?