We observed high DB request rates and identified a problem described in this upstream issue.
While the issue itself should be fixed upstream, the actual root cause might be originating in our host merging code.
- Given our host merge feature, is there a possibility that the host.rpc_seqno value of the host merged is set to the host_id of the merge target host, despite the fact that the target's host.rpc_seqno points back to the host being merged?
- Alternatively, does our host merge feature allow to set the host.rpc_seqno of the host merged to a host ID of a host that's also a "zombie" (userid==0)?
- Where can I find the host merge implementation?