tcp i/o timeouts connecting to databases
alexcurtin opened this issue · 3 comments
I am seeing intermittent timeouts in my network connecting to the database servers to do backups. We've eliminated DNS or conntrack issues and now I am left wondering if I can adjust the client timeout settings for initial connection to mongo?
I see the timeout settings available for backups operations, but can I configure the timeout on the connection to mongo or retries before failure? this is causing backups to fail mid backup, or before they begin in my case.
error getting count from db: connection() : dial tcp i/o timeout
im experiencing the same with my mongodb backups.
Also seeing this exact error with mgob v1.1 and 3-member replicaset on MongoDB 4.2.6 with readPreferred=secondary.
The Dockerfile for mongodb-tools is 4.2.1-r1
Current Alpine edge is 4.2.3-r1 https://pkgs.alpinelinux.org/packages?name=mongodb-tools&branch=edge&arch=x86_64
Current source release is 4.2.8 https://github.com/mongodb/mongo-tools/releases/tag/r4.2.8
I don't know if a tools upgrade would have any affect though.
For my case, the error was due to OOMKilled...increasing the resources memory limit for mgob resolved it.