-
Type: Bug
-
Resolution: Works as Designed
-
Priority: Minor - P4
-
None
-
Affects Version/s: 5.0.22
-
Component/s: Networking
-
None
-
Environment:Platform - GCP VM
CPU - x86
OS - Ubuntu 20.04.5 LTS
-
Networking & Observability
-
ALL
-
Security 2024-05-13, Security 2024-05-27, Security 2024-06-10, Security 2024-06-24, Networking & Obs 2024-09-16
We are using a sharded cluster of version 5.0.22. Recently, we have experienced that mongos reported the error "DNS resolution while connecting to peer was slow" for several times, which lasted about dozens of minutes each time. During this period, the cluster performance degraded significantly.
All shards and config servers hostname are added to /etc/hosts. We wonder why mongos still went to DNS instead of having got the address directly from the hosts file.
We find the Jira (https://jira.mongodb.org/browse/SERVER-53662) here with very similar symptom. According to the records, it linked to another one (https://jira.mongodb.org/browse/SERVER-59876). I do not know if they are related, but it doesn't look like the issue has been fixed.
Platform - GCP VM
CPU - x86
OS - Ubuntu 20.04.5 LTS
- related to
-
SERVER-94815 Improve the slow resolver message
- Backlog