Observed high CPU and disk I/O read with database unavailable #13374
Labels
Comments
|
@amanda-tian-awx Do you find any solution with disk and CPU usage on GCP + Rocky Linux? |
|
This log entry gives you a hint Please follow the Neo4j documentation about memory configuration. The lack of heap space is likely causing the GC pauses |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hey, I'm currently faced with a problem of database crash with very high CPU load and abnormal disk I/O multiple times, during which process observed
Detected VM stop-the-world pause.Configurations
Example bug report
Reproduce steps
Just installing the Neo4j database on the machine, loaded data via python driver and performed queries via driver or cypher.
Expected behavior: no database crash
Actual behavior
The problem has recurred 3 times during the past month, with interval of approximately 5-7 days. The downtime lasted for various of time, from hours to days with below observations:
Detected VM stop-the-world pause: {pauseTime=245207, gcTime=0, gcCount=0}in debug.logDuring the period when CPU peeks to 100%, I can't event ssh to the instance. The only way to recover was to restart the host machine. In the latest crash, though CPU recovered after 11 hours, database is still unavailable though
neo4j statusshows running status.I referred to the closed issue on
Detected VM stop-the-world pausedue to limited JVM memory. But this does resolve my problem on the continuous massive I/O operation and high CPU utilization. Any hints on what caused the situation without external heavy load of write/read operations?The text was updated successfully, but these errors were encountered: