Openstack MySQL Percona Cluster - too many connections in HA
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| mysql (Juju Charms Collection) |
New
|
Undecided
|
Unassigned | ||
Bug Description
Our HA setup looks like this:
Node 1-3 – Network / Most-Services-
Quantum-Gateway – Bare-Metal
RabbitMQ/MySQL/Ceph RADOS GW/Keystone/
Nodes 4-6 – Compute
Nova-Compute/Ceph – Bare-Metal
Nodes 7-8 – VSM (N1kv) Primary/Secondary
VSM – Bare-Metal
We were experiencing issues creating/deleting instances through horizon, and we also had slow connectivity through the dashboard. These issues came about after using the dashboard to test out dhcp in our networks, we were deleting and creating a lot of instances in a short period of time.
It seems that there were too many MySQL connections and some things we have done were:
-incremented the pool size
-intentionally minimized the amount of waiting time for connections
There are still a high amount of connections active at one time.
| tags: | added: cts |
