Support Questions

Find answers, ask questions, and share your expertise

Hive Downgrade

avatar

Hi all.

We would downgrad Hive 1.2.1 to 0.13.0 ( I mean, after an HDP 2.3 installation we have Hive 1.2.1).

Bacause , before with HDP 2.1 our job run fine and now, with HDP 2.3 we have errors.

We are on Productions and the fast way imho is a simple downgrade only for HIVE.

It's a good idea ? In order to downgrade I read :

Downgrade Hive

The Hive downgrade process reverses the upgrade process:

  • Downgrade WebHCat
  • Start the previous Hive Server 2 instance
  • Deregister the new Hive Server 2 instance
  • Downgrade Hive Metastore servers

When finished, validate the downgrade process.

It's correct ? Someone already tested ?

Thanks

1 ACCEPTED SOLUTION

avatar
Master Mentor

@Emanuele Parente Downgrade is not a good idea as you will lose tons of new features and performance improvements.

My suggestion is to work on fixing the issue instead downgrading.

View solution in original post

3 REPLIES 3

avatar
Master Mentor

@Emanuele Parente Downgrade is not a good idea as you will lose tons of new features and performance improvements.

My suggestion is to work on fixing the issue instead downgrading.

avatar
Master Mentor

@Emanuele Parente I suggest that you work with the support to address this issue.

avatar

Hi @Neeraj Sabharwal. Thanks for you reply.

The issue is related to

https://community.hortonworks.com/questions/18978/hive-121-issue-con-partitioned-table.html

But the main issue is on pre-existent query that when use partition field on WHERE condition cause an error. If the partition field is put in ON clause, the query (with complex left join) works fine.

About ACID transaction (ref. to https://community.hortonworks.com/questions/18978/hive-121-issue-con-partitioned-table.html ), I'm trying to set :

hive.txn.manager=org.apache.hadoop.hive.ql.lockmgr.DbTxnManager
hive.compactor.initiator.on=true
 hive.compactor.worker.threads=1

but also

hive.support.concurrency=true

hive.enforce.bucketing=true

hive.exec.dynamic.partition.mode=nonstrict

I hope that change something.

Regards,

Emanuele