Closed jbattiato closed 9 months ago
I've only tested the following clusters:
On two different storage classes:
The tablespace ones would not make sense for my use case.
Hardware:
Kubernetes:
etcd
workload: postgres
workload: pgbench
# Request Guaranteed QoS resources: requests: memory: '4Gi' cpu: 1.5 limits: memory: '4Gi' cpu: 1.5
# Memory configuration shared_buffers: '4GB' maintenance_work_mem: '1GB' effective_cache_size: '3GB'
Tests
I've only tested the following clusters:
On two different storage classes:
The tablespace ones would not make sense for my use case.
Environment
Hardware:
Kubernetes:
etcd
in high availability modeStorageClass
longhorn-pg:
localpath
Node Labels
workload: postgres
set in one of the master nodes with 8GB RAMworkload: pgbench
set in the Rpi with 4GB RAMCluster Parameters