logs under /var/log/hive not deleted

[#1]

we have ambari cluster with HIVE service

the ambari configuration required to delete the files under /var/log/hive , if files with the same name are more then 30

from HIVE config

hive_log_maxbackupindex=30

but when we access the master machines , we see more then 60 files with the same name

example:

cd /var/log/hive

ls -ltr | grep hivemetastore | wc -l

61

ls -ltr | grep hiveserver2 | wc -l

61

we also remove the remark from the line log4j.appender.DRFA.MaxBackupIndex , and restart the hive service

but this not help us

please advice what could be the problem ?

example of files under /var/log/hive

-rw-r--r--. 1 hive hadoop 2752 Sep 2 19:05 hivemetastore-report.json

-rw-r--r--. 1 hive hadoop 2756 Sep 2 19:05 hiveserver2-report.json

-rw-r--r--. 1 hive hadoop 636678 Sep 2 23:58 hiveserver2.log.2017-09-02

-rw-r--r--. 1 hive hadoop 1127874 Sep 2 23:59 hivemetastore.log.2017-09-02

-rw-r--r--. 1 hive hadoop 2369407 Sep 3 23:58 hiveserver2.log.2017-09-03

.

.

.

.

[#2]

@uri ben-ari

Can you please check your ambari-server UI to see if by any chance the hive-log4j has the following line Commented? (if yes, then please try to uncomment it

Ambari UI --> Hive --> Configs --> Advanced --> "Advanced hive-log4"

# 30-day backup
#log4j.appender.DRFA.MaxBackupIndex= {{hive_log_maxbackupindex}}

.

Once you uncomment the above section then try restarting the dependent services which shows "Restart Required" in ambari UI and then on the Hive server host please verify if the changes are reflected properly or not?

Example:

# grep 'MaxBackupIndex' /etc/hive/conf/hive-log4j.properties 
log4j.appender.DRFA.MaxBackupIndex= 30

.

[#3]

yes the parameter is - log4j.appender.DRFA.MaxBackupIndex=30 , and we already restart the hive service , but still files under /var/log/hive not deleted , what are the other checked that we need to do here ? , and what is the frequency that proccess need to delete the files ?