Hi Marko, have a look at Addressing random failures during NS8 Core updates. There can be some system directories with corrupted permissions.
Other users already experienced a similar problem: Cluster-admin not available after core update!
BTW, do you remember if you ran any of
- refresh/reload the cluster-admin page during the update
- press the Update Core button, then Cancel button
- press Update Core button on different browser windows or tabs
In other words, if any user action on the UI may have triggered the update core procedure twice in a short time?