CE part of https://github.com/hashicorp/vault-enterprise/pull/4269 #26406
Conversation
github-actions
bot
added
the
hashicorp-contributed-pr
|
CI Results: |
|
Build Results: |
| @@ -1682,11 +1682,17 @@ func (c *ServerCommand) Run(args []string) int { | |||
| level, err := loghelper.ParseLogLevel(config.LogLevel) | |||
| if err != nil { | |||
| c.logger.Error("unknown log level found on reload", "level", config.LogLevel) | |||
| goto RUNRELOADFUNCS | |||
| } else { | |||
There was a problem hiding this comment.
Arguably this should be fixed in its own PR, thoughts? The issue as I see it is that goto RUNRELOADFUNCS only makes sense when we're unable to re-read the config, because the reload steps we take after RUNRELOADFUNCS aren't exclusively dependent on the config file itself having changed - they rely on TLS certs changing or license file changes. But we have some cases where folks have copied the goto pattern thinking it's what should be done whenever any error occurs as part of reloading. This is undesirable: just because the user gave a bogus loglevel or whatnot, that shouldn't impede reloading other things like Census or the canary write interval.
There was a problem hiding this comment.
It seems fine to me to fix this in this PR, under that whole "make things nicer when you're in there" philosophy.
| @@ -1682,11 +1682,17 @@ func (c *ServerCommand) Run(args []string) int { | |||
| level, err := loghelper.ParseLogLevel(config.LogLevel) | |||
| if err != nil { | |||
| c.logger.Error("unknown log level found on reload", "level", config.LogLevel) | |||
| goto RUNRELOADFUNCS | |||
| } else { | |||
There was a problem hiding this comment.
It seems fine to me to fix this in this PR, under that whole "make things nicer when you're in there" philosophy.
No description provided.