Support Questions

Find answers, ask questions, and share your expertise
Announcements
Celebrating as our community reaches 100,000 members! Thank you!

You can't execute queries until the end of the 'atomic' block

avatar
Contributor

Hi,

I am getting above error from Hive while trying to query a table.

This is coincidentally during an insert operation on the table...

 

does it mean I cant access the table while a hive insert operation is ongoing...?

 

table contains lots of rows partitioned by two columns

 

 

2 ACCEPTED SOLUTIONS

avatar
Super Guru
Hi,

I don't see any errors posted in the post, can you please share again?

In general, Hive have locking mechanism enabled, so when INSERT is happening on a table, an exclusive lock will be place on the table and any query tries to run against the same table will have to wait, and there is a timeout to control how long the wait will be, and the query will fail with error if timeout exceeded, otherwise it will just keep waiting until lock is released.

So to understand more on the issue, I need to see what error was reported.

Cheers
Eric

View solution in original post

avatar
Contributor

Hi,

 

Issue resolved by increasing number of conection and session timeout as a workaround.

Actually need to use dedicated disk for  DB's.

 

Thanks,

Syed

View solution in original post

13 REPLIES 13

avatar
Super Guru
Hi,

I don't see any errors posted in the post, can you please share again?

In general, Hive have locking mechanism enabled, so when INSERT is happening on a table, an exclusive lock will be place on the table and any query tries to run against the same table will have to wait, and there is a timeout to control how long the wait will be, and the query will fail with error if timeout exceeded, otherwise it will just keep waiting until lock is released.

So to understand more on the issue, I need to see what error was reported.

Cheers
Eric

avatar
Contributor

Hi ,

 

im facing same issue , 

attached is the error screenshot


atomic error.jpg

avatar
Explorer

I am facing the same error when executing sql query through hue server。Now I have no idea to solve the error。 

 

Please。

avatar
Contributor

Hi,

 

Im also facing same issue ,

increased connections numbers to HUE DB, still issue exits. 

 

Thanks,

SS

avatar
Guru

Can you check "show locks" if the table is acquired EXCLUSIVE lock ?

avatar
Contributor

Hi,

 

No its not acquiring lock , result was empty set  in hue DB.

 

Thanks,

syed

avatar
Guru

DOes the same issue happen when you run via beeline ? I researched ,I found this occurs only in hue.

 

 

avatar
Contributor

Hi,

yes its happening at hue browser only

avatar
Expert Contributor

Hue is unable to handle the slow responses from Hive servers and hence the error. This happens when Hive is unable to return the query handle to Hue in a timely manner. So, the ideal solution here would be to investigate Hive slowness and not Hue.

 

Hue relies on server_conn_timeout parameter in its configuration (default 120s), after which the active sessions with hiveserver2 is teared down. 

 

If Hue is using MySQL as the underlying DB, then we can tweak the innodb-lock-wait-timeout to be greater than the server_conn_timeout in Hue's configuration, that way we get fail-fast system and the locked tables are released. Please note that this is only a workaround until the cause of Hive performance issue is found and remediated.