Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[repository-hdfs] java.io.IOException: Filesystem closed #174

Closed
x1a0 opened this Issue Mar 24, 2014 · 2 comments

Comments

Projects
None yet
2 participants
@x1a0
Copy link

commented Mar 24, 2014

Hi,

I am using elasticsearch-repository-hdfs 1.3.0.M2 with elasticsearch 1.0.1.
I have an embedded ES implementation. When it starts it registers a repository for snapshots on HDFS.

The very first time when it runs everything is OK. After I kill the process and run again, I got the exception pasted below. If I delete the data dir created by ES and run again. Everything goes back to well. So in short: I need to delete data to make it work.

If I use a repository on local filesystem there is no such problem.

The exception:

[info] 2014-03-24 13:12:38,144 ERROR c.m.p.Server$: org.elasticsearch.repositories.RepositoryException: [program] failed to list
snapshots in repository
[info] org.elasticsearch.repositories.RepositoryException: [program] failed to list snapshots in repository
[info]  at org.elasticsearch.repositories.blobstore.BlobStoreRepository.snapshots(BlobStoreRepository.java:350)
[info]  at org.elasticsearch.snapshots.SnapshotsService.snapshots(SnapshotsService.java:134)
[info]  at org.elasticsearch.action.admin.cluster.snapshots.get.TransportGetSnapshotsAction.masterOperation(TransportGetSnapshots
Action.java:91)
[info]  at org.elasticsearch.action.admin.cluster.snapshots.get.TransportGetSnapshotsAction.masterOperation(TransportGetSnapshots
Action.java:42)
[info]  at org.elasticsearch.action.support.master.TransportMasterNodeOperationAction$2.run(TransportMasterNodeOperationAction.ja
va:145)
[info]  at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
[info]  at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
[info]  at java.lang.Thread.run(Thread.java:744)
[info] Caused by: java.io.IOException: Filesystem closed
[info]  at org.apache.hadoop.hdfs.DFSClient.checkOpen(DFSClient.java:565)
[info]  at org.apache.hadoop.hdfs.DFSClient.listPaths(DFSClient.java:1524)
[info]  at org.apache.hadoop.hdfs.DFSClient.listPaths(DFSClient.java:1509)
[info]  at org.apache.hadoop.hdfs.DistributedFileSystem.listStatus(DistributedFileSystem.java:406)
[info]  at org.apache.hadoop.fs.FileSystem.listStatus(FileSystem.java:1462)
[info]  at org.apache.hadoop.fs.FileSystem.listStatus(FileSystem.java:1502)
[info]  at org.elasticsearch.hadoop.hdfs.blobstore.AbstractHdfsBlobContainer.listBlobsByPrefix(AbstractHdfsBlobContainer.java:93)
[info]  at org.elasticsearch.repositories.blobstore.BlobStoreRepository.snapshots(BlobStoreRepository.java:338)
[info]  ... 7 more

Thanks for outstanding project!

@costin

This comment has been minimized.

Copy link
Member

commented Mar 24, 2014

@zhangxiao Hi. The error indicates that you're using a Hadoop file-system which has been previously closed. The (HDFS) file system is closed at shutdown by the plugin which typically happens along side elasticsearch itself.

You mention that you are 'killing' elasticsearch - can you verify that the close is graceful? DistributedFileSystem has some files which are closed on process exit - if that doesn't occur (for whatever reason) then HDFS might act out.

I've tried replicating your use case but I couldn't. Also looking at the Hadoop code numbers it seems you are no running Hadoop 1.2.1 or 2.2.x

Can you confirm your Hadoop distribution, your plugin configuration and the basic actions to replicate this error? The smaller it is, the sooner I can replicate it.

Thanks!

@costin costin added bug labels May 2, 2014

@costin

This comment has been minimized.

Copy link
Member

commented May 2, 2014

Since there hasn't been any other update on this issue, I'm closing it for now. If it persists, please reopen it with the information requested above.

Thanks!

@costin costin closed this May 2, 2014

@costin costin added the invalid label May 2, 2014

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
You can’t perform that action at this time.