With the fix for deployed to Ignite, and both nodes started in rapid succession (the start was issued with just a few minutes in between, as opposed to the 10 minutes that I used before), we're now seeing cache inconsistencies in the outgoing server sessions.

Right after the restart, this seemed to affect pretty much every server session. After about half a day, only a few remain.

First node:

Second node:

It is interesting to see that using `netstat`, multiple outbound server connections to Dave's server (peirce.dave.cridland.net / 217.155.137.61) seem to have been established. Maybe the consistency check code is wrong to assume that there can be only one?

On first node:

On second node: