Cleanup and Maintenance
Use this when disk space is running low or when you need to regularly clear logs/temporary files.
In container mode, the logrotate inside the image (cronie + /etc/cron.daily/logrotate) automatically cleans up logs: daily / size>100M / 14-day gzip retention (for all 8 component logs). The UI "Temporary Cleanup" button uses a signal file
(/etc/kopens/host-action.request {"action":"tempClean"}) → host's kopens-host-action.path
triggers /data1/pp-temp/* clearing and container restart.
Docker image garbage collection: a weekly cron installed by install.sh (/etc/cron.weekly/plantpulse-cleanup)
automatically removes images older than 30 days / dangling layers / journald entries older than 30 days.
1. clean.sh — Delete Logs / Temporary Files in Bulk
$PE_HOME/bin/clean.sh
Cleanup targets (does not touch DB / user data — safe):
server/logs/*(Tomcat)db/logs/*(Cassandra log — sstable data remains intact)mqtt/log/*timeseries/*/log/*cache/log/*node/log/*/tmp/*
/tmp/* Bulk Cleanup — Removes Other Users' Temporary Files Tooclean.sh also clears system-wide /tmp. This removes temporary files from ssh session sockets / systemd-private / X11 / journal as well. Use with caution except on a gateway machine operated by a single user.
2. node-cleanup.sh — Reclaim Cassandra Disk Space
$PE_HOME/bin/node-cleanup.sh
To perform:
nodetool clearsnapshot --all— Reclaim disk space from outdated snapshotsnodetool cleanup— Remove unused sstables after token range changes
Normally, start.sh is automatically called in the background, but you can run it manually once more when disk space is low.
Takes a few minutes (proportional to the amount of sstables). Causes some I/O load — recommended during periods of low collection traffic.
3. Addressing Disk Exhaustion (Full Procedure)
# 1. 무엇이 차지 중?
df -h
du -sh $PE_HOME/*/log/* /data1/* 2>/dev/null | sort -h | tail -20
# 2. 로그 / 임시 파일 정리
$PE_HOME/bin/clean.sh
# 3. Cassandra 디스크 회수
$PE_HOME/bin/node-cleanup.sh
# 4. 오래된 백업 정리 (30 일 이상)
find /data1/pp-backup -maxdepth 1 -type d -mtime +30 -exec rm -rf {} +
# 5. (옵션) Docker 이미지 / 컨테이너 정리
docker system prune -af
# 6. 결과 확인
df -h
4. Common Pitfalls
| Symptom | Cause / Resolution |
|---|---|
ssh suddenly disconnects after clean.sh | The /tmp/* ssh socket was also deleted — reconnect via a new session |
Disk space not reclaimed even after node-cleanup.sh | sstable compaction in progress. Check progress with nodetool compactionstats |
find /data1/pp-backup -mtime +30 shows 0 items | Backups only exist for the last 30 days. Adjust the retention (e.g., -mtime +14) |
5. Further Reading
- Backup automation + retention periods: Backup (
backup.sh) - Disk diagnostics: Diagnostics / Inspection