Firstyear

well, there is also stuff related to cache behaviour and efficiency too that we could be capturing, but I just haven't exposed all of that either ... But yes, thanks...

@kellinm Yeah, in Australia the time to do a dup is about 5 times longer if we don't do mirrorsorcerer first. Do you know what the error reason was for...

Closed, I think we changed the dockerfile already

This really isn't a trivial issue. It's still a rust library but it just exposes a C compatible abi. So it would involve writing the needed compats and layers. As...

In non-release mode, just don't calculate the hashes at all then? Wouldn'th that resolve this?

To expand a bit more. 1. The client indicates intent to logout to a kanidm server 2. After the intent to logout is sent, the client removes it's local session...

> > the kanidm server processes the intent to logout (decision about sync vs async needed here ... probably sync) > > I'd say a sync processing of the logout...

> > sync vs replication delays > > Replication delay's always going to be a thing, and it's always part of the expectations management I guess. From a kill-(this|all)-tokens perspective...

I think it's a bit of a subtle difference. Once #159 has occured, and we reject the authentication, the account is in a way "locked". But that only locks that...

So how we would get the loglevel from config.log_level (https://github.com/kanidm/kanidm/blob/master/kanidmd/src/lib/core/mod.rs#L580 ) and get that into the right places?