Skip to content

Report an error for violation of serial keep policy#633

Open
tertsdiepraam wants to merge 1 commit into
mainfrom
keep-serial-error
Open

Report an error for violation of serial keep policy#633
tertsdiepraam wants to merge 1 commit into
mainfrom
keep-serial-error

Conversation

@tertsdiepraam
Copy link
Copy Markdown
Contributor

Closes #476

@Philip-NLnetLabs
Copy link
Copy Markdown
Member

This would be fine if the signing request is due to a new version of the zone. But signing requests also result from key management and timers. And I think we don't want an error for those cases.

@ximon18 ximon18 added this to the 0.1.0-beta1 milestone May 7, 2026
@ximon18
Copy link
Copy Markdown
Member

ximon18 commented May 11, 2026

This would be fine if the signing request is due to a new version of the zone. But signing requests also result from key management and timers. And I think we don't want an error for those cases.

I mistakenly set a zone to keep policy while also giving it a fast incremental resign rate which leads to Cascade repeatedly trying and failing to sign the zone due to keep policy violation, except that violation doesn't show up anywhere so I didn't realize what was going on. This is admittedly user error on my part, but this doesn't seem a useful state for Cascade to silently be in.

@Philip-NLnetLabs
Copy link
Copy Markdown
Member

If we make it not silent then we basically kill Keep because it will generate an error every time the refresh timer expires. Something at debug level is fine of course.

@tertsdiepraam
Copy link
Copy Markdown
Contributor Author

@Philip-NLnetLabs could you explain how you want this to work? I think the signer should not even start any resigning with keep. So we should add that to this PR maybe? Is there an easy way to do that?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

zone reload command without changing the SOA number can put cascade server into weird state.

3 participants