Skip to content

Node cannot recover when rejects term_unmatched AppendEntries #173

@akyriako

Description

@akyriako

node default_group:10.244.0.41:8107:8108 reject term_unmatched AppendEntries from 10.244.0.43:8107:8108 in term 916 prev_log_inde ││ x 44809 prev_log_term 916 local_prev_log_term 915 last_log_index 44809 entries_size 0 from_append_entries_cache: 0

  1. Pod is in Running state but internally typesense throws errors.
  2. The node will recover via the operator and elect a leader, after being terminated, but after some terms trying to replicate the log it will start throwing the same error again.
  3. The other pods are not affected
  4. Manually terminating the pod is fixing nothing.
  5. querying /status returns 401!!!

Interim solution: Manually deleting the PVC/PV and letting the Pod recreate the log from its peers solves the issue.

Metadata

Metadata

Assignees

Labels

bugSomething isn't working

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions