Receiving the error while creating the database with the same name which I dropper few seconds ago

I have dropped the database using below command.

DROP DATABASE IF EXISTS org2;

Try to recreate the database with the same name . Receiving the error.
create database org2;

Query execution failed

Reason:
SQL Error [1772] [HY000]: CREATE PARTITION DATABASE for operation CREATE DATABASE failed with error 1733:Leaf Error (10.100.0.18:3306): Unable to replay files for database ‘org2_6’. Run SHOW DATABASES EXTENDED and visit https://docs.memsql.com/docs/database#database-states for more information

What version are you on?

Send us a cluster report to bug-report@memsql.com

see here for how to take a report if you’ve never done that before:

I am facing the same error even when I create a new database. I thought it was because my leaves were down but ‘show leaves’ display all good.

My version is 7.0

ERROR 1772 (HY000): CREATE PARTITION DATABASE for operation CREATE DATABASE failed with error 1733:Leaf Error (:3306): Unable to replay files for database ‘dummy_1’. Run SHOW DATABASES EXTENDED and visit https://docs.memsql.com/docs/database#database-states for more information

On further analysis of logs I found following:

7108627380 2020-02-07 19:12:37.871 ERROR: Thread 99992: createDatabaseHelper: inquire: db creation fails because ‘MetadataManager::CreateDatabase’ failed with error 5.

I have also submitted my cluster report as an email.

So finally with help of @mikeczabator I was able to fix this issue. This error is very cryptic - better error reporting is needed. I was able to fix it as following

create database db1 partitions 4;

1 Like

Glad we got this figured out! For the sake of posterity, the issue @sanjeev.mishra hit was due to being out of disk space on the leaf servers. MemSQL 7.0+ pre-allocates log space for each partition - which is different from 6.x. A CREATE DATABASE command will fail if there is not enough disk available on the leaves for the pre-allocated logs.

Full details can be found here in the red box on top of the page : https://docs.memsql.com/v7.0/guides/cluster-management/operations/upgrading-to-70/

I opened a task to improve the error message around running out of disk during create database. Thanks for the feedback.