Configure environment variables¶

You can define custom environment variables to configure components in your Percona XtraDB Cluster. This is useful for customizing HAProxy settings, adding PMM Client options, or configuring other cluster components.

The Operator stores environment variables in Kubernetes Secrets . You can then reference these Secrets in your cluster configuration.

Configure HAProxy environment variables¶

This example shows how to set up environment variables for HAProxy configuration:

apiVersion: v1
kind: Secret
metadata:
  name: my-env-var-secrets
type: Opaque
data:
  HA_CONNECTION_TIMEOUT: MTAwMA==
  OK_IF_DONOR: MQ==
  HA_SERVER_OPTIONS: Y2hlY2sgaW50ZXIgMzAwMDAgcmlzZSAxIGZhbGwgNSB3ZWlnaHQgMQ==
  PEER_LIST_SRV_PROTOCOL: dGNw

Note

The variables in this example have the following effects:

HA_CONNECTION_TIMEOUT sets a custom timeout for HAProxy health checks. HAProxy repeatedly executes status queries on XtraDB Cluster instances. The default 10-second timeout works for most workloads, but you may need to increase it for unstable Kubernetes networks or when soft lockups occur on Kubernetes nodes.
OK_IF_DONOR allows application connections to XtraDB Cluster donor nodes. Donor nodes run backups, which can slow down SQL queries. Enable this option when only one XtraDB Cluster node is available and a second node is joining the cluster via SST.
HA_SERVER_OPTIONS sets custom options for servers in the HAProxy configuration file. The default is check inter 30000 rise 1 fall 5 weight 1. You can add additional options referenced in the HAProxy documentation .
PEER_LIST_SRV_PROTOCOL enables you define what protocol (UDP or TCP) the Operator uses when performing peer-list SRV lookups. The use of TCP may be required for large database clusters with many nodes where peer-list SRV lookup returns large DNS responses or when UDP DNS queries are blocked by network policies. You can configure the protocol for HAProxy or ProxySQL.

Create the Secret¶

Environment variables are stored as base64-encoded strings in the data section. You need to encode each variable value.

For example, to set HA_CONNECTION_TIMEOUT to 1000, run this command:

$ echo -n "1000" | base64 --wrap=0

Note

On Apple macOS, use this command instead:

$ echo -n "1000" | base64

You can also decode base64-encoded values to verify them:

$ echo "MTAwMA==" | base64 --decode
1000

Apply the configuration¶

Create the Secret with the following command:

$ kubectl create -f deploy/my-env-secret.yaml

Add the Secret name to your cluster configuration. Edit the deploy/cr.yaml file and add the envVarsSecret key to the appropriate section: pxc, haproxy or proxysql. The sample configuration for HAproxy is:
```
haproxy:
  envVarsSecret: my-env-var-secrets
```
Apply the updated configuration:
```
$ kubectl apply -f deploy/cr.yaml
```

Configure alternative memory allocator¶

You can use an alternative memory allocator library for mysqld to optimize memory usage. This is often recommended when memory usage is higher than expected.

The Percona XtraDB Cluster Pods include the jemalloc allocator. You can enable it with the LD_PRELOAD environment variable:

LD_PRELOAD=/usr/lib64/libjemalloc.so.1

Here’s how:

Create a Secret with the encoded value. This is the example of the Secret’s YAML manifest:

apiVersion: v1
kind: Secret
metadata:
  name: my-new-env-var-secrets
type: Opaque
data:
  LD_PRELOAD: L3Vzci9saWI2NC9saWJqZW1hbGxvYy5zby4x

Create the Secret object:

$ kubectl create -f deploy/my-new-env-var-secret.yaml

Add the Secret to the PXC section in your deploy/cr.yaml file:
```
pxc:
  envVarsSecret: my-new-env-var-secrets
```
Apply the configuration:
```
$ kubectl apply -f deploy/cr.yaml
```

Last update: 2025-08-14