Common administrative tasks

lorem ipsum

Stop, restart, and revive

This procedure can be useful in cases where you want to change AWS instances to save money. For example, you can use us-east instances during the day and switch to us-west instances at night.

This example uses the database created by the following command using the region us-east-1:

$ vcluster create_db --db-name test_db --hosts 192.2.0.1,192.2.0.2,192.2.0.3 --catalog-path /scratch_b/qa --data-path /scratch_b/qa --shard-count 4 --communal-storage-location s3://testbucket/test_db --depot-path /path/to/depot --depot-size 20G --config /opt/vertica/config/vertica_cluster.yaml --config-param awsauth=key:secret,awsenablehttps=0,awsregion=us-east-1,awsendpoint=myhost:9000 --password "" --skip-package-install
✔ Check NMA service health
...
✔ Synchronize catalog with communal storage
[INFO] Successfully created a database with name [test_db]

  1. Stop the database:

    $ Stop DB
/opt/vertica/bin/vcluster stop_db --db-name test_db --config /opt/vertica/config/vertica_cluster.yaml --password ""
✔ Collect node information
✔ Collect cluster information
✔ Update node state from running database
✔ Collect information for all up nodes
✔ Synchronize catalog with communal storage
✔ Stop database
✔ Verify database is not running
[INFO] Successfully stopped a database with name test_db

  2. Revive the database. For this example, the database is revived into a different region:

    $ vcluster revive_db --db-name test_db --hosts 192.2.0.1,192.2.0.2,192.2.0.3 --communal-storage-location s3://testbucket/test_db --config /opt/vertica/config/vertica_cluster.yaml --config-param awsauth=key:secret,awsenablehttps=0,awsregion=us-west-1,awsendpoint=myhost:9000
✔ Check NMA service health
✔ Verify database is running
✔ Download cluster_config.json
✔ Create necessary directories on Vertica hosts
✔ Get network profile of cluster
✔ Load remote catalog
[INFO] Successfully revived database test_db

  3. Start the database:

    $ /opt/vertica/bin/vcluster start_db --db-name test_db --hosts 192.2.0.1,192.2.0.2,192.2.0.3 --catalog-path /path/to/catalog --config /opt/vertica/config/vertica_cluster.yaml --config-param awsauth=miniokey:miniosecret,awsenablehttps=0,awsregion=us-east-1,awsendpoint=qastress-39:9000 --password ""
✔ Check NMA service health
✔ Collect nodes information
✔ Download cluster_config.json
✔ Check NMA service health
✔ Verify database is running
✔ Read catalog
✔ Check Vertica version
✔ Get contents of vertica.conf
✔ Get contents of spread.conf
✔ Start 3 node(s)
✔ Wait for 3 node(s) to come up: all nodes are up
✔ Synchronize catalog with communal storage
✔ Collect node information
✔ Collect cluster information
✔ Update node state from running database
[INFO] Started database test_db

Test on sandboxed subclusters

You can create sandboxed subclusters and perform tests on them without affecting your production database. For details, see Subcluster sandboxing:

  1. Add the subcluster sc1 which contains nodes 192.2.0.4 and 192.2.0.5:

    $ vcluster add_subcluster  --subcluster sc1 --db-name test_db --password "" --hosts 192.2.0.1,192.2.0.2,192.2.0.3 --control-set-size 1 --new-hosts 192.2.0.4,192.2.0.5
✔ Collect cluster information
✔ Check NMA service health
...
✔ Initiate rebalance of subcluster shards
[INFO] Successfully added subcluster sc1 with nodes [192.2.0.3,192.2.0.4] to database **test_db**

  2. Sandbox the subcluster sc1 with a the new sandbox sand:

    $ vcluster sandbox_subcluster --subcluster sc1 --sandbox sand -p "" --config /opt/vertica/config/vertica_cluster.yaml
✔ Collect information for all up nodes
✔ Find all subclusters and record their sandboxing information
✔ Convert subcluster into sandbox in catalog system
✔ Wait for subcluster nodes to come up
[INFO] Successfully sandboxed subcluster sc1 as sand

  3. Verify that your nodes were sandboxed with list_all_nodes. The following command is run from outside the sandbox sand, so the state of nodes in sand are listed as UNKNOWN:

    $ vcluster list_all_nodes --config /opt/vertica/config/vertica_cluster.yaml -p ""
✔ Collect node information
✔ Collect cluster information
✔ Update node state from running database
✔ Check NMA service health
✔ Read Vertica version
✔ Check node state from running database
[
{
"address": "192.0.2.1",
"name": "v_test_db_node0001",
"state": "UP",
"catalog_path": "/scratch_b/qa/test_db/v_test_db_node0001_catalog/Catalog",
"subcluster": "default_subcluster",
"sandbox": "",
"is_primary": true,
"version": "v24.3.0-20240613"
},
{
"address": "192.0.2.2",
"name": "v_test_db_node0002",
"state": "UP",
"catalog_path": "/scratch_b/qa/test_db/v_test_db_node0002_catalog/Catalog",
"subcluster": "default_subcluster",
"sandbox": "",
"is_primary": true,
"version": "v24.3.0-20240613"
},
{
"address": "192.0.2.3",
"name": "v_test_db_node0003",
"state": "UP",
"catalog_path": "/scratch_b/qa/test_db/v_test_db_node0003_catalog/Catalog",
"subcluster": "default_subcluster",
"sandbox": "",
"is_primary": true,
"version": "v24.3.0-20240613"
},
{
"address": "192.0.2.4",
"name": "v_test_db_node0004",
"state": "UNKNOWN",
"catalog_path": "/scratch_b/qa/test_db/v_test_db_node0004_catalog/Catalog",
"subcluster": "sc1",
"sandbox": "sand",
"is_primary": false,
"version": "v24.3.0-20240613"
},
{
"address": "192.0.2.5",
"name": "v_test_db_node0005",
"state": "UNKNOWN",
"catalog_path": "/scratch_b/qa/test_db/v_test_db_node0005_catalog/Catalog",
"subcluster": "sc1",
"sandbox": "sand",
"is_primary": false,
"version": "v24.3.0-20240613"
}
]
[INFO] Successfully listed all nodes

  4. After you finish testing your sandboxed subcluster, unsandbox it:

    $ vcluster unsandbox_subcluster --subcluster sc1 -p "" --config /opt/vertica/config/vertica_cluster.yaml
✔ Collect node information
✔ Collect cluster information
✔ Update node state from running database
✔ Check NMA service health
✔ Collect information for all up nodes
✔ Stop node
✔ Wait for subcluster nodes to come down
✔ Convert sandboxed subcluster into regular subcluster in catalog
✔ Delete database directories
✔ Check Vertica version
✔ Get Vertica startup command for unsandboxed nodes
✔ Start 0 node(s)
✔ Wait for subcluster nodes to come up
[INFO] Successfully unsandboxed subcluster sc1