hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jean-Daniel Cryans <jdcry...@apache.org>
Subject Re: Couple of interesting observations for Hbase 0.90.4.
Date Thu, 20 Oct 2011 20:35:41 GMT
This is a discussion that comes every once in a while, not sure if we have a
FAQ entry or in the book...

But anyways, short answer is putting your computer to sleep is not

Long answer is that putting your computer to sleep is like partitioning that
node from the rest of the cluster, so the zookeeper timeout still applies
and if it sleeps longer than 3 minutes (by default) than it's considered
dead. The fact that you have only one node is an unfortunate event because
it means nothing else can failover the components that are running there.

We could have a special case for standalone setups, but beyond that I don't
see it being easily feasible.


On Wed, Oct 19, 2011 at 7:39 PM, Jignesh Patel <jigneshmpatel@gmail.com>wrote:

> 1. Whenever my computer goes to sleep mode(Mac lion OS X) it stops
> HMaster.However all map reduce related not remain active.
> 739 JobTracker
> 585 DataNode
> 675 SecondaryNameNode
> 496 NameNode
> 5991 Jps
> 826 TaskTracker
> Is this expected behaviour?
> 2. Even HMaster is not running when I tried to check the status - instead
> of
> giving appropriate error it gives
> ERROR: org.apache.hadoop.hbase.ZooKeeperConnectionException: HBase is able
> to connect to ZooKeeper but the connection closes immediately. This could
> be
> a sign that the server has too many connections (30 is the default).
> Consider inspecting your ZK server logs for that error and then make sure
> you are reusing HBaseConfiguration as often as you can. See HTable's
> javadoc
> for more information.
> -Jignesh

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message