0

When running a simple example I get this error.

I tried changing permissions and used different directories.

Caused by: java.io.IOException: FailedToCheckpointException(message:Failed to rename /home/bryan/tachyon-0.7.1/underFSStorage/tmp/tachyon/workers/1445054000001/3/8 to /home/bryan/tachyon-0.7.1/underFSStorage/tmp/tachyon/data/8)
    at tachyon.worker.WorkerClient.addCheckpoint(WorkerClient.java:130)
    at tachyon.client.TachyonFS.addCheckpoint(TachyonFS.java:228)
    at tachyon.client.FileOutStream.close(FileOutStream.java:105)

Any ideas?

My answer below does not actually solve this issue. It's only a workaround using HDFS as the UnderFS.

My config here uses the local file system:

export TACHYON_UNDERFS_ADDRESS=$TACHYON_HOME/underFSStorage
dtolnay
  • 9,621
  • 5
  • 41
  • 62
BAR
  • 15,909
  • 27
  • 97
  • 185
  • This is a similar issue. Please try it and let me know if the workaround fix your issue: https://groups.google.com/forum/#!topic/tachyon-users/yNY0YflcLvA – Kamran Oct 17 '15 at 04:56

1 Answers1

0

I got past this error by first checking my Hadoop version:

hadoop version
> Hadoop 2.7.1

Then compiling Tachyon for Hadoop 2.7.0: (for 2.7.1 it would not compile)

mvn -Dhadoop.version=2.7.0 clean install

And setting TACHYON_UNDERFS_ADDRESS to a local HDFS address:

nano conf/tachyon-env.sh
export TACHYON_UNDERFS_ADDRESS=hdfs://my-cluster-m:8020

This leads to another problem which might affect you (if running single node):

Tachyon on Dataproc Master Replication Error

Community
  • 1
  • 1
BAR
  • 15,909
  • 27
  • 97
  • 185