[ 
https://issues.apache.org/jira/browse/HBASE-1948?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Gautam Sampathkumar updated HBASE-1948:
---------------------------------------


Hi Stack,
          I think this problem is seen regardless of the size of the table, but 
I have a table that's 'column-heavy'. I have 2 rows and about 10K+ columns per 
row. Database size is about 25MB.

No instance of HBase region server is started up because I have 
hbase.cluster.distributed set to false. I have Hbase set to manage my Zookeeper 
instance.

I believe that the problem may be related to the fact that hadoop's namenode 
may rearrange block information when it starts up and this is not reflected in 
hbase which still attempts to use old locations.

I hope that helps. Please let me know if you need more information. I will 
attach my hdfs-site.xml and hbase-site.xml. Since this is my test/dev 
environment the settings are pretty minimal

> updates to existing tables do not get reflected after a complete HBase + 
> Hadoop + Zookeeper restart
> ---------------------------------------------------------------------------------------------------
>
>                 Key: HBASE-1948
>                 URL: https://issues.apache.org/jira/browse/HBASE-1948
>             Project: Hadoop HBase
>          Issue Type: Bug
>          Components: client
>    Affects Versions: 0.20.1
>         Environment: ubuntu 9.04
> hadoop 0.20.1 (single machine mode)
> hbase 0.20.0 (pseudo-distributed mode i.e. using HDFS on a single machine)
> zookeeper 3.2.1
>            Reporter: Gautam Sampathkumar
>
> I have an existing hbase table with data in it. When I restart hadoop, hbase 
> and zookeeper, I am able to read all of the data that existed in the table 
> prior to the restart. However when I write data to the table, it does not get 
> reflected. 
> When I do a disable <table> and then an enable <table> on an hbase shell. The 
> data that was written to the table now appears and the table is up to date. 
> However, even after this my thrift client still sees the old data and not the 
> updated values.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to