Chalk this one up to ignorance. I did some code tracing and learned that the /accumulo/instance_id is actually a directory in hdfs. The sole file in that directory is named after the instance id.
$ hadoop fs -ls /accumulo/instance_id Found 1 items -rw-r--r-- 1 medined supergroup 0 2012-06-11 17:58 /accumulo/instance_id/b519799c-3a51-4c9b-af21-96d577e2c11f My system might be more stable now that I have Zookeeper storing information someplace other than /tmp. On Sun, Jun 10, 2012 at 9:58 PM, David Medinets <[email protected]> wrote: > I deleted /accumulo from HDFS. I ran "ls /" in Zookeeper and did not > see /accumulo. So I ran the 'accumulo init' program. No errors. > > Then I ran ZooZap with the verbose option and got this error: > > 12/06/10 21:57:06 ERROR client.ZooKeeperInstance: unable obtain > instance id at /accumulo/instance_id > Exception in thread "main" java.lang.RuntimeException: Accumulo not > initialized, there is no instance id at /accumulo/instance_id > at > org.apache.accumulo.core.client.ZooKeeperInstance.getInstanceIDFromHdfs(ZooKeeperInstance.java:263) > at > org.apache.accumulo.server.client.HdfsZooInstance._getInstanceID(HdfsZooInstance.java:123) > at > org.apache.accumulo.server.client.HdfsZooInstance.getInstanceID(HdfsZooInstance.java:116) > at org.apache.accumulo.server.util.ZooZap.main(ZooZap.java:67) > > Has something changed in how the accumulo information is stored in Zookeeper? > > Looking at Zookeeper, I now see: > > [zk: 127.0.0.1:2181(CONNECTED) 1] ls / > [accumulo, zookeeper] > [zk: 127.0.0.1:2181(CONNECTED) 2] ls /accumulo > [2b64e012-a197-4ec0-a487-dc2e154cbd94, instances] > [zk: 127.0.0.1:2181(CONNECTED) 3] ls /accumulo/instances > [development] > [zk: 127.0.0.1:2181(CONNECTED) 4] ls /accumulo/instances/development > [] > [zk: 127.0.0.1:2181(CONNECTED) 5] ls > /accumulo/2b64e012-a197-4ec0-a487-dc2e154cbd94 > [tables, tservers, users, problems, next_file, root_tablet, > hdfs_reservations, tracers, gc, table_locks, config, masters]
