Added extra option to add readOnly thrift HMS uri #308
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…d on read only calls for better spread of traffic
patduin
commented
Jan 23, 2024
waggle-dance-api/src/main/java/com/hotels/bdp/waggledance/api/model/AbstractMetaStore.java
Show resolved
Hide resolved
patduin
commented
Jan 23, 2024
...core/src/main/java/com/hotels/bdp/waggledance/client/SplitTrafficMetastoreClientFactory.java
Show resolved
Hide resolved
jmnunezizu
requested changes
Feb 8, 2024
README.md
Outdated
| @@ -167,6 +167,7 @@ The table below describes all the available configuration values for Waggle Danc | |||
| | `primary-meta-store.hive-metastore-filter-hook` | No | Name of the class which implements the `MetaStoreFilterHook` interface from Hive. This allows a metastore filter hook to be applied to the corresponding Hive metastore calls. Can be configured with the `configuration-properties` specified in the `waggle-dance-server.yml` configuration. They will be added in the HiveConf object that is given to the constructor of the `MetaStoreFilterHook` implementation you provide. | | |||
| | `primary-meta-store.database-name-mapping` | No | BiDirectional Map of database names and mapped name, where key=`<database name as known in the primary metastore>` and value=`<name that should be shown to a client>`. See the [Database Name Mapping](#database-name-mapping) section.| | |||
| | `primary-meta-store.glue-config` | No | Can be used instead of `remote-meta-store-uris` to federate to an AWS Glue Catalog ([AWS Glue](https://docs.aws.amazon.com/glue/index.html). See the [Federate to AWS Glue Catalog](#federate-to-aws-glue-catalog) section.| | |||
| | `primary-meta-store.read-only-remote-meta-store-uris` | No | Can be used to configure an extra read-only endpoint for the primary Metastore. This is an optimization if your environment runs separate Metastore endpoints and traffic needs to be differted efficiently. Waggle Dance will direct traffic to the read-write or read-only endpoints based on the call being done. For instance `get_table` will be a read-only call but `alter_table` will be forwarded to the read-write Metastore.| | |||
| if (metaStore.getReadOnlyRemoteMetaStoreUris() != null) { | ||
| CloseableThriftHiveMetastoreIface readWrite = newHiveInstance(metaStore, name, metaStore.getRemoteMetaStoreUris(), | ||
| properties); | ||
| CloseableThriftHiveMetastoreIface readOnly = newHiveInstance(metaStore, name+"_ro", |
...core/src/main/java/com/hotels/bdp/waggledance/client/SplitTrafficMetastoreClientFactory.java
Show resolved
Hide resolved
jmnunezizu
reviewed
Feb 8, 2024
...core/src/main/java/com/hotels/bdp/waggledance/client/SplitTrafficMetastoreClientFactory.java
Show resolved
Hide resolved
jmnunezizu
approved these changes
Feb 8, 2024
githubjianli
approved these changes
Feb 8, 2024
mroark1m
approved these changes
Feb 8, 2024
README.md
Outdated
| @@ -167,6 +167,7 @@ The table below describes all the available configuration values for Waggle Danc | |||
| | `primary-meta-store.hive-metastore-filter-hook` | No | Name of the class which implements the `MetaStoreFilterHook` interface from Hive. This allows a metastore filter hook to be applied to the corresponding Hive metastore calls. Can be configured with the `configuration-properties` specified in the `waggle-dance-server.yml` configuration. They will be added in the HiveConf object that is given to the constructor of the `MetaStoreFilterHook` implementation you provide. | | |||
| | `primary-meta-store.database-name-mapping` | No | BiDirectional Map of database names and mapped name, where key=`<database name as known in the primary metastore>` and value=`<name that should be shown to a client>`. See the [Database Name Mapping](#database-name-mapping) section.| | |||
| | `primary-meta-store.glue-config` | No | Can be used instead of `remote-meta-store-uris` to federate to an AWS Glue Catalog ([AWS Glue](https://docs.aws.amazon.com/glue/index.html). See the [Federate to AWS Glue Catalog](#federate-to-aws-glue-catalog) section.| | |||
| | `primary-meta-store.read-only-remote-meta-store-uris` | No | Can be used to configure an extra read-only endpoint for the primary Metastore. This is an optimization if your environment runs separate Metastore endpoints and traffic needs to be differted efficiently. Waggle Dance will direct traffic to the read-write or read-only endpoints based on the call being done. For instance `get_table` will be a read-only call but `alter_table` will be forwarded to the read-write Metastore.| | |||
flaming-archer
added a commit
to flaming-archer/waggle-dance
that referenced
this pull request
Jun 26, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
tldr; Split traffic based on called HMS API method, e.g.
getTablewill go to a readOnly HMS andalterTablewill go to readWrite HMSThe problem addressed here is running WD at scale. Generally our company deploys Waggle Dance as part of an Apiary Data lake: https://github.com/ExpediaGroup/apiary-data-lake.
This involves deploying ReadOnly and ReadWrite Metastores (HMS).
For the primary (local) metastore waggle dance is configured to the ReadWrite instance which connects to a ReadWrite RDS backend. This means all traffic both read and writes end up on our ReadWrite RDS instance. This PR tries to split that traffic and move read traffic to ReadOnly instance.
The benefit would be: