[10:17] <stickupkid> manadart, https://github.com/juju/juju/pull/12196
[10:50] <Hybrid512> Hi
[10:51] <Hybrid512>  I had a juju upgrade this morning (2.8.6/snap) which went well.
[10:51] <Hybrid512> I then upgraded the juju controller and as soon as I did that, it disturbed my previous deployments
[10:51] <Hybrid512> agent was unavailable (which is quite normal since controller was upgrading itself) but then, when it went up again, it looks like the model was reconfiguring itself for every charms which broke some VIPs and had quite some impacts
[10:52] <Hybrid512> agent was unavailable (which is quite normal since controller was upgrading itself) but then, when it went up again, it looks like the model was reconfiguring itself for every charms which broke some VIPs and had quite some impacts
[10:52] <Hybrid512> I didn't thought the controller would be that critical for **deployed** applications ... I could easily imagine that it would be critical to **new** deployments but not already deployed ones
[10:52] <Hybrid512> what is the best practice in such case ?
[11:05] <stickupkid> Hybrid512, firstly I believe a bug at the very least would be good to track this, as it does sound like we did something wrong
[11:06] <Hybrid512> stickupkid I can create it ... just didn't know if this was a bug or if I did something wrong
[11:06] <Hybrid512> so at fisrt, would be great to know your recommended best practices towards that
[11:07] <stickupkid> I don't know enough, maybe manadart does
[11:09] <manadart> Hybrid512: What version was the upgrade from?
[11:10] <Hybrid512> 2.8.5 -> 2.8.6
[11:10] <Hybrid512> from snap package
[15:33] <stickupkid> hml, https://github.com/juju/juju/pull/12232
[15:33] <hml> stickupkid: ack