[SRU] Puppet master fails with 'stack level too deep' error when storeconfigs = true

[Lionel Porcheron]

[Impact]

* puppet setup with storeconfig activated

* On client run, master fail with "stack level too deep" error, and client do no apply any of its manifests

* not all clients are affected, but if one client is failing it will fail forever

* This issue is known and fixed on upsteam, https://tickets.puppetlabs.com/browse/PUP-1064

[Test Case]

To reproduce this issue, you need a puppet master with storeconfig=true running on Ubuntu Trusty.

Apply the rules with "puppet agent -t"

Without fix, you may face an "stack level too deep", with fix, manifests should be applied as expected.

[Regression Potential]

Minimal, fix is upstream and applied in new puppet versions.

On our side, we have validated the fix since once week on a > 500 machines setup and this error does not occur anymore.

[Lionel Porcheron (lionel.porcheron)]

Attached debdiff integrating upstream fix for trusty.

Issue is fixed in Utopic (puppet 3.5.1)

[Mathew Hodson (mhodson)]

Should have been fixed in Utopic and later. Upstream bug says fixed in PUP 3.5.0.

[Robie Basak (racb)]

Thank you for the patch, and sorry for the delay. It looks like this wasn't in the sponsorship queue, which accounts for most of the delay. Thank you to Mathew for noticing and adding it.

We're making an effort to clear out languishing items in the sponsorship queue today, so I'm tackling this now. It looks like Markus did correctly follow the SRU procedure at the time (except to subscribe ~ubuntu-sponsors, hence the delay).

But we will still need someone to perform SRU verification to make an upload worthwhile. Is there someone willing to volunteer to do this, or does this bug not matter to anyone any more?

If someone is willing, please resubscribe ~ubuntu-sponsors to the bug and we can upload. I'll subscribe myself to the bug so we can address it promptly this time.

[Robie Basak (racb)]

For reference: https://wiki.ubuntu.com/StableReleaseUpdates