🦭 kubectl-mdba

kubectl-mdba (mariadb administrator) is a kubectl plugin that helps manage mariadb-operator async/semi-sync replication clusters

Disclaimer

• This project is not affiliated with MariaDB Foundation or MariaDB Operator in any way.
• It has not been verified by MariaDB Operator, and no representations are made regarding the quality of the upstream software.
• Intended for mariadb-operator v26.3.0

Dependencies

• kubectl - https://github.com/kubernetes/kubectl
• jq - https://github.com/jqlang/jq

Optionally, you can install the following tools to improve the output of some commands:

• bat - https://github.com/sharkdp/bat
• yq - https://github.com/mikefarah/yq (or https://github.com/kislyuk/yq)

Installation

Krew

❯ kubectl krew install mdba

Manual

Place kubectl-mdba into the directory within your $PATH (e.g. ~/.local/bin)

Completion

Place kubectl_complete-mdba into the directory within your $PATH (e.g. ~/.local/bin)

Customization

You can export the following variables to tweak the plugin's behaviour.

VARIABLE	DEFAULT	DETAILS
KMDBA_KUBECTL	kubectl	kubectl binary name
KMDBA_KUBECONFIG	$KUBECONFIG	default kubeconfig path
KMDBA_NAMESPACE	default	default k8s namespace
KMDBA_RECREATE_STEP	1	initial step number from 1 to 14 for replica recreation
KMDBA_BACKUP_THREADS	1	number of threads from 1 to 16 to use for parallel datafiles transfer
KMDBA_STREAM_PORT	4444	default port from 1024 to 65535 for backup stream
KMDBA_SKIP_CONFIRMATION	0	skip confirmation
KMDBA_IGNORE_PRIMARY_MISMATCH	0	ignore mismatched desired and current primary pod indexes
KMDBA_BACKUP_DIR	/var/lib/mysql/.kmdba_backup	tmp directory in /var/lib/mysql/ to accept backup stream
KMDBA_RESTORE_DIR	/var/lib/mysql/.kmdba_restore	tmp directory in /var/lib/mysql/ for restored backup

Usage

kubectl mdba helps manage mariadb-operator async/semi-sync replication clusters

Usage:
  kubectl mdba <command> [<target>]

Commands:
  ls [<mariadb>]            list mariadbs with their pods
  status <mariadb>          check mariadb status
  suspend <mariadb>         pause mariadb reconciliation
  unsuspend <mariadb>       resume mariadb reconciliation
  enter <pod>               exec into pod
  sql <pod>                 launch mariadb shell
  proc <pod>                print processlist
  du <pod> [-d <db>]        calculate database disk usage
  top [<pod>]               display cpu and ram usage
  repl <replica>            check replication status
  skip <replica>            skip erroneous transactions
  prom <replica>            promote replica to primary
  recreate <replica>        recreate replica from primary

Flags:
  -h, --help                show this message
  -v, --version             show plugin version
  -b, --kubectl <bin>       set kubectl binary
  -k, --kubeconfig <file>   set kubeconfig path
  -n, --namespace <ns>      set namespace scope
  -y, --yes                 skip confirmation

Recreate Flags:
  -f, --force               ignore primary index mismatch
  -p, --port <num>          port for backup stream (default: 4444)
  -t, --threads <num>       parallel threads {1..16} for datafiles transfer (default: 1)
  -s, --step <num>          step {1..16} to start from (default: 1)
  -g, --gtid <gtid>         set GTID when --step is between 10 and 14

Links

• https://mariadb.com/kb/en/setting-up-a-replica-with-mariabackup
• mariadb-operator/mariadb-operator#141