How to fix a cluster startup failure due to meta election timeout in a single meta node setup?
I encountered a cluster startup failure after upgrading to version 1.5. The meta node log showed that the election of the election client was being blocked. We have only one meta node in the cluster. How can I fix this issue by adjusting the meta_leader_lease_secs value?
André Falk
Asked on Dec 11, 2023
-
Identify the issue: The cluster startup failure was caused by the meta election timeout in a single meta node setup.
-
Adjust meta_leader_lease_secs: Increase the timeout of the meta election by adjusting the meta_leader_lease_secs value to a larger value, such as 3600.
-
Impact on the cluster: Since there is only one meta node, increasing the meta_leader_lease_secs timeout won't impact the cluster negatively.
-
Resolution: By adjusting the meta_leader_lease_secs value, the meta election timeout issue can be resolved, allowing the cluster to start successfully.
-
Confirmation: After adjusting the meta_leader_lease_secs value, verify that the cluster starts without any issues.