You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hbase.apache.org by pola sai ram <sa...@gmail.com> on 2018/07/23 11:18:00 UTC

Data Mismatch for a row in HBase

Hello Everyone,

I am currently facing *data* *mismatch* for a *single row* out of millions
of rows in *Primary* and *Replicated* (Master-Master) clusters. *No writes* for
same rowkey happened on both clusters at *same time*. There is *No
replication lag*. There is *No clock skew* on both the clusters.
Below is the mismatched row:

*Primary Cluster:  *
(Some Row Key)-> column=0:(XYZ), timestamp=*1530252316292*, *value*
=\x80\x00\x01dJ%G\*x87*\x00\x00\x00\x00
Here \x80\x00\x01dJ%G\x87\x00\x00\x00\x00 = 2018-06-29 06:05:*14.503*

*Replicated Cluster:*
(Same Row Key)-> column=0:(XYZ), timestamp=*1530252316292*, *value*
=\x80\x00\x01dJ%G\*xFC*\x00\x00\x00\x00
Here \x80\x00\x01dJ%G\xFC\x00\x00\x00\x00 = 2018-06-29 06:05:*14.620 *

One of the *possible* scenario is that *write* has happened at *same
timestamp* for *same rowKey* but with *different data* on both the clusters.
 But we found that it is *very unlikely to happen *in our case
Can there be any other scenarios for which this could have happened ? We
are using *0.98.24* version of HBase

Thanks,
Sairam Pola