Today, some staging deployments failed and I found out one of the servers has run out of disk space, where /var/lib/docker/containers had grown unexpectedly large because of large log files.
Finding the largest folders:
1# start with /2sudo du -h -x --max-depth=1 / | sort -hr3 4# then go into largest one one by one5sudo du -h -x --max-depth=1 /var | sort -hr6sudo du -h -x --max-depth=1 /var/lib | sort -hr7sudo du -h -x --max-depth=1 /var/lib/docker | sort -hrThen, I found /var/lib/docker/containers was the culprit.
This server ran multiple docker containers, none of which were particularly disk-intensive. GPT suggested this could mostly be related to large log file size. To check:
1sudo du -sh /var/lib/docker/containers/*/*.logSure enought, there was a huge 140GB log file from clickhouse.
To truncate all log files,
1sudo truncate -s 0 /var/lib/docker/containers/*/*.logTo limit the log file size, you can add a /etc/docker/daemon.json file with the following container to use JSON logging with a limit.
1{2 "log-driver": "json-file",3 "log-opts": {4 "max-size": "10m",5 "max-file": "3"6 }7}Make sure to restart docker (sudo systemctl restart docker) to apply changes.
Comments