Thanks Sean.
Later I will open a JIRA for this issue. Now I am not very clear when this problem will happen, it always along with the hbase region server thread crashed(on my three node cluster, all regionservers died).
BTW, How do you usually load your data into Trafodion besides using hive, typically when the data quantity is large?

Sean Broeder <sean.broeder@esgyn.com>于2016年9月13日周二 下午9:27写道:

Hi Qiao,

I am very sorry you ran into this issue.  I’ve never seen this before, but I don’t typically use hive to load my tables.  I think you should open a JIRA for this. Do you know if the problem is repeatable?  If so, please put any details to reproduce into the JIRA.  Also, if you could set log4j.logger.org.apache.hadoop.hbase.coprocessor.transactional=trace  in hbase/conf /log4j.properties that would be helpful.

 

Thanks,

Sean

 

From: 乔彦克 [mailto:qyanke@gmail.com]
Sent: Tuesday, September 13, 2016 1:37 AM
To: dev@trafodion.incubator.apache.org; user@trafodion.incubator.apache.org


Subject: trafodion get stuck and hbase log shows "followed by a smaller key"

 

 

Hi, all,

    I use hive to load some data to do some test. This morning after I loaded some data I did a sum query which leading to a HBase regionserver crashed. Then I restart HBase hoping to do some load balace, but just got these errors in the HBase regionserve log. 

    The worse is that I have no other choice but to execute "initialize trafodion, drop" because the trafci can only execute this command.

image001.jpg

 

image002.jpg

 

   some other Trafodion table region has the same problems.

   Do you ever encountered this problems?

   Any replies is appreciated.

 

  Thanks,

  Qiao