Upgrade to Version 3.7.x

Prerequisite

Upgrade the Docker version to 20.10.x or above.

Steps to upgrade the Docker version:

1. Uninstall the current Docker:

yum remove -y docker-ce docker-ce-cli containerd.io

2. Install the latest version of Docker:

yum install -y docker-ce docker-ce-cli containerd.io

• To install a specific version of Docker:

yum install -y docker-ce-24.0.5 docker-ce-cli-24.0.5 containerd.io

• Replace 24.0.5 with the appropriate version string
• To list specific versions of docker

yum list docker-ce --showduplicates | sort -r

3. Start Docker:

systemctl start docker

Migrations Steps from Version 3.6.x to 3.7.0

Pulse Upgrade Steps (Core Services Server / Addon Services Server)

1. Update Configuration:

   • Update ImageTag in the $AcceloHome/config/accelo.yml file to 3.7.0.

2. Push Updated Configuration to Database:

   • Run the following command:

accelo admin database push-config

3. Ensure that you have downloaded and loaded all the necessary docker images for 3.7.0.
4. Download a new accelo binary and replace the existing one with the new accelo binary.

5. Run the migration command. For example:
   1. -v 3.7.0 : The version you are migrating to.
   2. -i 3.6.2 : The version you are migrating from.
   3. To migrate from the version 3.6.2 to 3.7.0, the migration command looks as shown below.

For a single cluster:

accelo migrate -v 3.7.0 -i 3.6.2 -b

For multi-cluster:

accelo migrate -v 3.7.0 -i 3.6.2 -a -b

6. Run the same command again, If any issue occurs during migration on any step.
7. If YARN optimizer is deployed, append the feature flag "enableStaticQueue": true to the ad-graphql section in the ad-core.yml file.

vi $AcceloHome/config/docker/ad-core.yml

After adding the mentioned feature flag, the FEATURE_FLAG environment must look like this. If you have additional FEATURE_FLAG entries, do not need to remove them, just append "enableStaticQueue": true at the end . Example:

'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": "", "enableStaticQueue" : true }'

8. Restart all the containers so that they use the updated images.

accelo restart all -d

9. Uninstall the Hydra Agent.

accelo uninstall remote

10. Install the Hydra Agent.

accelo deploy hydra

11. If Logsearch is deployed in standalone mode, verify whether the yarn app es_urls parameter is present in the override.yml file. If it exists, update the port number from 19013 to 19051.

yarn_app_es_urls: <http://<LOG_SEARCH_HOSTNAME>>:19051

12. Reconfigure the cluster to get the updated configuration for Agents.

accelo reconfig cluster -a

13. To push the updated configuration to database, execute the following command.

accelo admin database push-config -a

14. To configure when gauntlet snap must run the MongoDB cleanup and compaction, run the following command.

accelo config retention

15. Finally execute the following command to create indices for database collections.

If readOnlyRootFSEnabled is set to true in $AcceloHome/config/accelo.yml, then follow the steps listed below instead of running the command directly.

​

1. Open $AcceloHome/config/accelo.yml and set the value of readOnlyRootFSEnabled to false.

2. Run accelo admin database push-config 

3. Run accelo restart all -d.

4. Run accelo admin database index-db.

5. Open $AcceloHome/config/accelo.yml and set the value of readOnlyRootFSEnabled to true.

6. Run accelo admin database push-config.

7. Run accelo restart all -d.

accelo admin database index-db

16. If the older playbooks already present in the node where ad-director is deployed, delete them, and download the latest playbooks tar from the License UI.
    1. Delete the older playbooks using the command: rm -rf $AcceloHome/work/<cluster_name>/director/playbook/*.Replace <cluster_name> with name of the cluster where the playbooks are present.
    2. Download the playbooks tar from the License UI. For details, see Deploy Playbooks.

Migration Steps from Version 3.5.x to 3.7.0

Pulse Upgrade Steps (Core Services Server / Addon Services Server)

1. Update Configuration:

   • Update ImageTag in the $AcceloHome/config/accelo.yml file to 3.7.0.

2. Push Updated Configuration to Database:

   • Run the following command:

accelo admin database push-config

3. Ensure that you have downloaded and loaded all the necessary docker images for 3.7.0.
4. Download a new accelo binary and replace the existing one with the new accelo binary.

5. Run the migration command. For example:
   1. -v 3.7.0 : The version you are migrating to.
   2. -i 3.5.2 : The version you are migrating from.
   3. To migrate from the version 3.5.2 to 3.7.0, the migration command looks as shown below.

For a single cluster:

accelo migrate -v 3.7.0 -i 3.5.2 -b

For multi-cluster:

accelo migrate -v 3.7.0 -i 3.5.2 -a -b

6. Run the same command again, If any issue occurs during migration on any step.
7. Delete the following directories, if they are re present.

rm -rf $AcceloHome/data/pg/*

rm -rf $AcceloHome/data/pg-ui/*

8. Delete the metric groups from the alerts collection in MongoDB.

Log into mongo db and run following commands.

Note: For details about logging into MongoDB, reach out to Acceldata Support.

use alerts

db.metric_groups.deleteMany({})

db.metrics.deleteMany({})

9. If YARN optimizer is deployed, append the feature flag "enableStaticQueue": true to the ad-graphql section in the ad-core.yml file.

After adding the mentioned feature flag, the FEATURE_FLAG environment must look like this. If you have additional FEATURE_FLAG entries, do not need to remove them, just append "enableStaticQueue": true at the end . Example:

vi $AcceloHome/config/docker/ad-core.yml

'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": "", "enableStaticQueue" : true }'

10. Restart all the containers so that they use the updated images.

accelo restart all -d

11. Uninstall the Hydra Agent.

accelo uninstall remote

12. Install the Hydra Agent.

accelo deploy hydra

13. If Logsearch is deployed in standalone mode, verify whether the yarn app es_urls parameter is present in the override.yml file. If it exists, update the port number from 19013 to 19051.

yarn_app_es_urls: <http://<LOG_SEARCH_HOSTNAME>>:19051

14. Reconfigure the cluster to get the updated configuration for Agents.

accelo reconfig cluster -a

15. To push the updated configuration to database, execute the following command.

accelo admin database push-config -a

16. To configure when gauntlet snap must run the MongoDB cleanup and compaction, run the following command.

accelo config retention

17. Finally execute the following command to create indices for database collections.

If readOnlyRootFSEnabled is set to true in $AcceloHome/config/accelo.yml, then follow the steps listed below instead of running the command directly.

​

1. Open $AcceloHome/config/accelo.yml and set the value of readOnlyRootFSEnabled to false.

2. Run accelo admin database push-config 

3. Run accelo restart all -d.

4. Run accelo admin database index-db.

5. Open $AcceloHome/config/accelo.yml and set the value of readOnlyRootFSEnabled to true.

6. Run accelo admin database push-config.

7. Run accelo restart all -d.

accelo admin database index-db

18. If the older playbooks already present in the node where ad-director is deployed, delete them, and download the latest playbooks tar from the License UI.
    1. Delete the older playbooks using the command: rm -rf $AcceloHome/work/<cluster_name>/director/playbook/*.Replace <cluster_name> with name of the cluster where the playbooks are present.
    2. Download the playbooks tar from the License UI. For details, see Deploy Playbooks.