Pratyush Bhatt created HDDS-10920:
-------------------------------------
Summary: [Hbase-Ozone] Too many RITs on Hbase restart
Key: HDDS-10920
URL: https://issues.apache.org/jira/browse/HDDS-10920
Project: Apache Ozone
Issue Type: Bug
Reporter: Pratyush Bhatt
Attachments: image-2024-05-28-15-57-19-952.png,
image-2024-05-28-16-00-49-366.png, image-2024-05-28-16-10-04-866.png
When HBase is set to run on top of Ozone, and we restart HBase which had some
tables and data written into it earlier, can notice there are so many Regions
in Transitions for a long duration, and many operations either fails or take
abnormally huge amount of time, so we have to wait until all the RITs are
gone/reduced to a small number.
!image-2024-05-28-15-57-19-952.png|width=788,height=233! Can see here approx 2k
regions are in transition, and I tried doing a simple table create command with
some Column families and pre-splitting to distribute the regions even on the
RS's(there were 9 in my cluster so approx 10 regions pre RS), it took more than
10 minutes.
{code:java}
hbase:005:0> create 'rittable', 'cf1','cf2','cf3','cf4','cf5', SPLITS =>
(1..90).map {|i| "#{i.to_s.rjust(2, '0')}"}
ERROR: The procedure 78012 is still running
For usage try 'help "create"'Took 613.8492 seconds
hbase:006:0> {code}
From HBase UI this was shown for the procedure shown in ERROR.
!image-2024-05-28-16-00-49-366.png|width=1027,height=67!
Matter of fact, tried same thing without any pre-splitting, this command also
took 10+ minutes.
{noformat}
hbase:016:0>
hbase:017:0> create 'rittable2', 'cf1','cf2','cf3','cf4','cf5'
ERROR: The procedure 80054 is still running
For usage try 'help "create"'
Took 608.6806 seconds{noformat}
!image-2024-05-28-16-10-04-866.png|width=872,height=52!
cc: [~weichiu] [~Sammi] [~ashishk]
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
