Hive Metastore aborted transaction loop

[#1]

Hello

Just upgraded to HDP 2.5.3 from HDP 2.3.6 and experiencing lots of problems. One for instance is that the 3 Hive metastores that I've running for HA sake are printing this to the metastore.log like crazy every few milliseconds:

2017-02-16 13:58:01,341 INFO  [org.apache.hadoop.hive.ql.txn.AcidHouseKeeperService-0]: txn.TxnHandler (TxnHandler.java:performTimeOuts(2960)) - Aborted 0 transactions due to timeout
2017-02-16 13:58:01,345 INFO  [org.apache.hadoop.hive.ql.txn.AcidHouseKeeperService-0]: txn.TxnHandler (TxnHandler.java:performTimeOuts(2949)) - Aborted the following transactions due to timeout: []
2017-02-16 13:58:01,345 INFO  [org.apache.hadoop.hive.ql.txn.AcidHouseKeeperService-0]: txn.TxnHandler (TxnHandler.java:performTimeOuts(2960)) - Aborted 0 transactions due to timeout
2017-02-16 13:58:01,350 INFO  [org.apache.hadoop.hive.ql.txn.AcidHouseKeeperService-0]: txn.TxnHandler (TxnHandler.java:performTimeOuts(2949)) - Aborted the following transactions due to timeout: []
2017-02-16 13:58:01,350 INFO  [org.apache.hadoop.hive.ql.txn.AcidHouseKeeperService-0]: txn.TxnHandler (TxnHandler.java:performTimeOuts(2960)) - Aborted 0 transactions due to timeout
2017-02-16 13:58:01,354 INFO  [org.apache.hadoop.hive.ql.txn.AcidHouseKeeperService-0]: txn.TxnHandler (TxnHandler.java:performTimeOuts(2949)) - Aborted the following transactions due to timeout: []
2017-02-16 13:58:01,354 INFO  [org.apache.hadoop.hive.ql.txn.AcidHouseKeeperService-0]: txn.TxnHandler (TxnHandler.java:performTimeOuts(2960)) - Aborted 0 transactions due to timeout
2017-02-16 13:58:01,359 INFO  [org.apache.hadoop.hive.ql.txn.AcidHouseKeeperService-0]: txn.TxnHandler (TxnHandler.java:performTimeOuts(2949)) - Aborted the following transactions due to timeout: []

I cannot find anything pointing to the cause of this problem. Any hint?

[#2 · Accepted solution]

This has been fixed in https://issues.apache.org/jira/browse/HIVE-15181. In the meantime you can

set hive.direct.sql.max.query.length=1;set 

hive.direct.sql.max.elements.in.clause=1000;

for the Standalone Metastore process(es).

[#3]

Thanks! It worked! I put the in Ambari in custom-hive-site, restarted the affected services and no more noise.