TDSQL
TMA-TDSQL-12.7
Ensure ZooKeeper Orphan Node Cleanup Is Configured
Description
TDSQL creates many task/alert/state ZNodes in ZooKeeper. Abnormal exits or network jitter can produce orphan nodes, and long-term accumulation causes: ZK memory bloat, affecting overall availability Residual alert/task nodes leaking historical sensitive information Historical credential/password fields remaining readable long-term TDSQL provides two cleanup mechanisms: <clear_zknode value="15"/> - cleans up ordinary ...
Méthode de vérification
Méthode d'audit: Revue manuelle
# Check scheduler.xml grep -E '(clear_zknode|alarm_zknode)' /path/to/tdsql/conf/scheduler.xml # Check for residual historical nodes in ZK (should not exceed recent counts) zkCli.sh -server <zk_ip>:<port> ls /tdsql/job ls /tdsql/alarm
Méthode de correction
Méthode de correction: Recommandation ou correctif d'infrastructure
<!-- Enable orphan node cleanup (retain 15 days) --> <clear_zknode value="15"/> <!-- Enable alert node cleanup (24-hour expiration, checked hourly) --> <alarm_zknode open_clear="1" expire_time="86400" check_interval="3600"/>
Propriétés
- Systèmes d'exploitation
- Tous
- Étiquettes
- tdsqlzookeeper
- Référence
- CIS Tencent TDSQL 10.3.22 Benchmark v1.0.0 - 12.7 Ensure ZooKeeper Orphan Node Cleanup Is Configured