Hello Avery,
That most definitely helps. Thank you!
Steve
From: Avery Ching [mailto:ach...@apache.org]
Sent: Tuesday, March 13, 2012 9:30 PM
To: giraph-user@incubator.apache.org
Subject: Re: Zookeeper errors when running tests.
I believe those are safe. It's just letting us know that a node was already
created. We should probably try to reduce the amount of messages in the
future. ZooKeeper is automatically instantiated for you if you do not set the
zookeeper location.
In GiraphJob, you can set the ZooKeeper location:
/**
* Utilize an existing ZooKeeper service. If this is not set, ZooKeeper
* will be dynamically started by Giraph for this job.
*
* @param serverList Comma separated list of servers and ports
*(i.e. zk1:2221,zk2:2221)
*/
public final void setZooKeeperConfiguration(String serverList) {
conf.set(ZOOKEEPER_LIST, serverList);
}
Or you can set at runtime with "giraph.zkList".
Hope that helps,
Avery
On 3/13/12 1:23 PM, Fleischman, Stephen (ISS SCI - Plano TX) wrote:
When running the giraph test, I'm seeing Zookeeper errors, or at least warnings:
12/03/13 14:06:57 INFO server.PrepRequestProcessor: Got user-level
KeeperException when processing sessionid:0x1360d74bff2 type:create
cxid:0x25 z= xid:0xfffe txntype:unknown reqpath:n/a Error
Path:/_hadoopBsp/job_local_0001 Error:KeeperErrorCode =3D NodeExists for
/_hadoopBsp/job_loca=
l_0001
Is this significant? The jobs seem to run.
Where is Zookeeper configured? I would like to be using a replicated Zookeeper
environment.
Thank you in advance,
Steve Fleischman