- Subscribe to RSS Feed
- Mark Question as New
- Mark Question as Read
- Float this Question for Current User
- Bookmark
- Subscribe
- Mute
- Printer Friendly Page
Maximum hive.exec.max.dynamic.partitions allowed & recommended
- Labels:
-
Apache Hadoop
-
Apache Hive
Created 06-15-2017 08:29 PM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi Team, any suggestions. what will be the impact of adding this setting hive.exec.max.dynamic.partitions to whitelist. which allows any user to create any number of partitions for a table at run time.
Created 06-15-2017 09:52 PM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
There is no maximum as per my knowledge and again this value depends on the back-end metastore database what you are using.
I have tested up 500,000 in production with oracle as back-end.
hive.exec.max.dynamic.partitions=500000
There won't be any impact by adding that to whitelist but always suggested to have number so that it won't impact cluster in long term.
Example: One user is keeps on increasing partitions, where each partition is very small file, in this case it increase the Namenode metadata which proportionally my effect cluster.
Created 06-15-2017 09:52 PM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
There is no maximum as per my knowledge and again this value depends on the back-end metastore database what you are using.
I have tested up 500,000 in production with oracle as back-end.
hive.exec.max.dynamic.partitions=500000
There won't be any impact by adding that to whitelist but always suggested to have number so that it won't impact cluster in long term.
Example: One user is keeps on increasing partitions, where each partition is very small file, in this case it increase the Namenode metadata which proportionally my effect cluster.
Created 06-16-2017 03:31 PM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
We are using oracle as back-end, we have plan to implement it in production in near future. As you said small files will increase metadata which proportionally effects cluster. So, are there any precautions or suggestion to be followed while implementing or after implementation.
