sql: avoid tenant ID reuse #100615
Labels
A-multitenancy
Related to multi-tenancy
branch-master
Failures and bugs on the master branch.
branch-release-23.1
Used to mark GA and release blockers, technical advisories, and bugs for 23.1
C-bug
Code not up to spec/doc, specs & docs deemed correct. Solution expected to change code/behavior.
GA-blocker
T-multitenant
Issues owned by the multi-tenant virtual team
Tenant ID reuse is known to incur correctness problems in the following two cases:
we do not yet wait for server shutdown when the service mode is changed to 'NONE'. This makes it possible to start dropping a tenant and then reusing its ID before servers from the previous record have fully shut down. This creates a possibility for a past server to start serving traffic for a freshly created tenant, which is unacceptable.
we have a cache of tenant capabilites on every node. We do not yet implement an invalidation protocol to clear that cache upon tenant deletion. This makes it possible for requests from a new tenant to be authorized using the capabilities of a previous tenant, which is also unacceptable.
Until we implement the necessary synchronization, we can avoid the correctness issues by preventing ID reuse.
Jira issue: CRDB-26522
The text was updated successfully, but these errors were encountered: