Scalding Spyglass HBase connection issue

17 views
Skip to first unread message

Deepika Khera

unread,
Jun 27, 2015, 1:33:23 PM6/27/15
to scaldi...@googlegroups.com
Hi,

I am trying to use spyglass to do a scan from scalding to hbase.

class SimpleHBaseSourceExample(args: Args) extends JobBase(args) with HBasePipeConversions {

   val output = args("output")


   val hbs = new HBaseSource(

     "deepika_test",

    "zookeeperhost:2181",

     new Fields("key"),

     List("cf"),

     List(new Fields("name")),

     sourceMode = SourceMode.GET_LIST, keyList = List("1", "2", "3"))

     .read

     .fromBytesWritable(new Fields("key", "name"))

     .write(Tsv(output format "get_list"))


 }


When I run my job, it seems to be able to connect to zookeeper, but then stalls for the time. At the end it throws an exception. Please see logs below:


15/06/26 23:05:34 INFO zookeeper.ClientCnxn: Session establishment complete on server zookeeperhost/1.2.3.4:2181, sessionid = 0x24db7c727c200a2, negotiated timeout = 30000

15/06/26 23:06:02 INFO util.Update: newer Cascading release available: 2.6.3

15/06/26 23:16:14 ERROR hbase.HBaseInputFormatBase: ************* HBase table deepika_test is not accessible

15/06/26 23:16:14 ERROR hbase.HBaseInputFormatBase: org.apache.hadoop.hbase.client.NoServerForRegionException: Unable to find region for deepika_test,,99999999999999 after 10 tries.

at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation.locateRegionInMeta(HConnectionManager.java:980)

at org.apache.hadoop.hbase.client.HConnectionManager$HConnectionImplementation.locateRegion(HConnectionManager.java:885)


The zookeeper logs at this time show the following errors:


2015-06-26 23:08:01,649 - INFO  [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@1007] - Closed socket connection for client /1.2.3.4:42460 (no session established for client)

2015-06-26 23:09:01,661 - INFO  [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxnFactory@197] - Accepted socket connection from /1.2.3.4:42515

2015-06-26 23:09:01,662 - WARN  [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxn@357] - caught end of stream exception

EndOfStreamException: Unable to read additional data from client sessionid 0x0, likely client has closed socket

        at org.apache.zookeeper.server.NIOServerCnxn.doIO(NIOServerCnxn.java:228)

        at org.apache.zookeeper.server.NIOServerCnxnFactory.run(NIOServerCnxnFactory.java:208)

        at java.lang.Thread.run(Thread.java:745)



Would appreciate any help or pointers on this?


Thanks,
Deepika

Asher

unread,
Sep 3, 2015, 9:18:01 AM9/3/15
to Scalding Development
Hi Deepika,

Are you able to do a full scan on the table and issue get commands via the hbase shell for the keys in your list?
It doesn't seem like a scalding issue, but more a zookeeper/hbase connectivity issue.  You might have better luck in an HBase group for that.
--Asher
Reply all
Reply to author
Forward
0 new messages