High Availability for GraphQL

The objective of the High availability (HA) is to have the Pulse User Interface (UI) on a parallel node so that if one of the nodes goes down, the other node would be available and all the traffic can be diverted to it. This feature is available on Pulse version 1.7.0 onwards.

User handles the Additional Load Balancers to automatically distribute the workloads.

Prerequisites: The Accelo Utility must already be installed by the client and must be an active node. To install the Accelo Utility check the Prerequisites. For first time installation of Accelo Utility contact support@acceldata.io
Illustration: Load Balancer

Illustration: Load Balancer

Steps to Deploy High Availability

Follow the steps to deploy the HA for GraphQL.

Step 1: Amend the Active Node

Step 2: Amend the HA Node Post Deployment

First Host - Amend the Active Node

Execute the following steps on the Active node.

  1. Replace the ports and hostname values for influx, elastic and vm sections of the following services in the <ACCELO_HOME>config/acceldata_<cluster_name>.conf file as described in the following table:

Replace the docker container-names, with the hostname on which the influx-db,__elastic, vm-select and vm-insert are deployed.

You must change the ports from the internal docker ports to the ports exposed on the active host. host6 has been used here as an example. Use the correct host name as per the your requirement.
BeforeAfter

influx = [

{

name = "default"

host = "ad-tsdb"

port = 8086

username = "<username>"

password = "<password>"

dbName = "<db_name>"

},

{

name = "kafka"

host = "ad-tsdb"

port = 8086

username = "<username>"

password = "<password>"

dbName = "<db_name>"

},

{

name = "graphite"

host = "ad-tsdb"

port = 8086

username = "<username>"

password = "<password>"

dbName = "<db_name>"

},

{

name = "nifi"

host = "ad-tsdb"

port = 8086

username = "<username>"

password = "<password>"

dbName = "<db_name>"

},

{

name = "dsConnection",

host = "ad-tsdb",

port = 8086,

username = "<username>"

password = "<password>"

encrypted = false,

dbName = "<db_name">

}

],

elastic = [

{

name = "default"

host = "ad-elastic"

port = 9200

},

{

name = "fsanalytics"

host = "ad-elastic"

port = 9200

},

{

name = "nifi"

host = "ad-elastic"

port = 9200

}

]

vm = [

{

name = "default",

writeURL = "http://ad-vminsert:8480/insert/3609009724/influx",

readURL = "http://ad-vmselect:8481/select/3609009724/prometheus",

dbName = "ad_cdp710_demo",

clusterNameHash = "3609009724"

}

influx = [

{

name = "default"

host = "host6"

port = 19009

username = "<username>"

password = "<password>"

dbName = "<db_name>"

},

{

name = "kafka"

host = "host6"

port = 19009

username = "<username>"

password = "<password>"

dbName = "<db_name>"

},

{

name = "graphite"

host = "host6"

port = 19009

username = "<username>"

password = "<password>"

dbName = "<db_name>"

},

{

name = "nifi"

host = "host6"

port = 19009

username = "<username>"

password = "<password>"

dbName = "<db_name>"

},

{

name = "dsConnection",

host = "host6",

port = 19009,

username = "<username>"

password = "<password>"

encrypted = false,

dbName = "<db_name>"

}

],

elastic = [

{

name = "default"

host = "host6"

port = 19013

},

{

name = "fsanalytics"

host = "host6"

port = 19013

},

{

name = "nifi"

host = "host6"

port = 19013

}

]

vm = [

{

name = "default",

writeURL = "http://host6:19043/insert/3609009724/influx",

readURL = "http://host6:19042/select/3609009724/prometheus",

dbName = "ad_cdp710_demo",

clusterNameHash = "3609009724"

}
  1. Deploy the core, and type 'Y' in all prompts:
Bash
Copy
  1. Now, restart all core components. The changes are now saved in the active node.
Bash
Copy

Second Host - Amend the HA Node Post Deployment

  1. Docker Installation
  2. Pulse CLI download and initialization
  3. Create license file under work directory and copy the contents from first host work/license
  4. In the CLI, populate the configuration yaml file by running the following command: accelo admin makeconfig ad-ha-graphql. This displays a prompt asking to overwrite the config file.
  5. Type 'Y' to confirm.
  6. After the file is generated, perform the following changes highlighted in bold:

Add Line 17 and 18 if you are a Cloudera platform user.

Add the host and ports property for Alerts and Actions in the environment section.
BeforeAfter

version: "2"

services:

ad-ha-graphql:

image: ad-graphql

container_name: ad-ha-graphql

environment:

MONGO_URI=ZN4v8cuUTXYvdnDJIDp+R8Z+ZsVXXjv8zDOvh8UwQXrj0RTboRglzSSmtH8Gu0LC

MONGO_ENCRYPTED=true

MONGO_SECRET=Ah+MqxeIjflxE8u+/wcqWA==

UI_PORT=4000

LDAP_HOST=ad-ldap

LDAP_PORT=19020

DS_HOST=ad-query-estimation

DS_PORT=8181

'FEATURE_FLAGS={ "ui_regex": { "regex": "ip-([^.]+)", "index": 1 }, "rename_nav_labels":{}, "timezone": "", "experimental": true, "themes": false, "hive_const":{ "HIVE_QUERY_COST_ENABLED": false, "HIVE_MEMORY_GBHOUR_COST": 0, "HIVE_VCORE_HOUR_COST": 0 }, "spark_const": { "SPARK_QUERY_COST_ENABLED": false, "SPARK_MEMORY_GBHOUR_COST": 0, "SPARK_VCORE_HOUR_COST": 0 }, "queryRecommendations": false, "hostIsTrialORLocalhost": false, "data_temp_string": "" }'

volumes:

/etc/localtime:/etc/localtime:ro

/data01/acceldata/work/license:/etc/acceldata/license:ro

ulimits: {}

ports: 4000:4000

depends_on: []

opts: {}

restart: ""

extra_hosts: []

network_alias: []

label: HA GraphQL

version: "2"

services:

ad-ha-graphql:

image: ad-graphql

container_name: ad-ha-graphql

environment:

MONGO_URI=ZN4v8cuUTXYvdnDJIDp+R8Z+ZsVXXjv8zDOvh8UwQXrj0RTboRglzSSmtH8Gu0LC

MONGO_ENCRYPTED=true

MONGO_SECRET=Ah+MqxeIjflxE8u+/wcqWA==

UI_PORT=4000

LDAP_HOST=ad-ldap

LDAP_PORT=19020

DS_HOST=ad-query-estimation

DS_PORT=8181

DB_HOST=<DB_HOST_IP>

DB_PORT=19030

CAP_SCHEDULER_ENABLED_YARN = false

PLATFORM = CLDR #(HWX for HDP clusters)

DASHPLOT_IP=<DASHPLOT_IP>

DASHPLOT_PORT=18080

ALERTS_IP=<ALERTS_IP>

ACTIONS_IP=<ACTIONS_IP>

ALERTS_PORT=19015

ACTIONS_PORT=19016

'FEATURE_FLAGS={ "ui_regex": { "regex": "ip-([^.]+)", "index": 1 }, "rename_nav_labels":{}, "timezone": "", "experimental": true, "themes": false, "hive_const":{ "HIVE_QUERY_COST_ENABLED": false, "HIVE_MEMORY_GBHOUR_COST": 0, "HIVE_VCORE_HOUR_COST": 0 }, "spark_const": { "SPARK_QUERY_COST_ENABLED": false, "SPARK_MEMORY_GBHOUR_COST": 0, "SPARK_VCORE_HOUR_COST": 0 }, "queryRecommendations": false, "hostIsTrialORLocalhost": false, "data_temp_string": "" }'

volumes:

/etc/localtime:/etc/localtime:ro

/data01/acceldata/config/hosts:/etc/hosts:ro

/data01/acceldata/work/license:/etc/acceldata/license:ro

ulimits: {}

ports:

4000:4000

depends_on: []

opts: {}

restart: ""

extra_hosts: []

network_alias: []

label: HA GraphQL

If the user has LDAP and Proxy enabled on the Active node, it must be mirrored in the HA node.

  1. For optional services, ad-ldap follow LDAP integration similar to first host and update config/ldap/ldap.conf

  2. For optional services, ad-proxy follow same steps for Pulse UI TLS integration (placing certs file and removing 4000 port from graphql yaml ) and one additional step for ad-ha-graphql

    1. Update url mentioning ad-graphql to ad-ha-graphql in file config/proxy/config.toml

  3. Update either /etc/hosts entries or create a separate host file at location config/host to add following host entry:

Bash
Copy
  1. Optional: In case of using config/host file, update the volumes file mount in config/docker/addons/ad-ha-graphql.yml :
Bash
Copy
  1. Deploy respective addons to start UI service on second host, by selecting the ad-graphql, ad-ldap(optional), and ad-proxy(optional) services. ad-ha-graphql is now deployed:

The deployment of High availability for GraphQL is now complete.
Type to search, ESC to discard
Type to search, ESC to discard
Type to search, ESC to discard
Next to read:
Standalone Logsearch

For additional help, contact www.acceldata.force.com OR call our service desk +1 844 9433282

Copyright © 2025

On This Page
High Availability for GraphQLSteps to Deploy High AvailabilityFirst Host - Amend the Active NodeSecond Host - Amend the HA Node Post Deployment