Menu

#445 Copy to Repository - out of space revert

Next Release
open
High
2017-04-20
2017-04-20
No

When transferring a job with just tiff images (7.1.0 release) the Copy to Repository step resulted in the following error:

Server Error: A directory for this job already exists (/repository/repository02/2017/04201233). Please contact the system administrator to manually resolve this issue.

The error in the server logs was as follows:

20 Apr 2017 12:43:38,437 [Thread-2] ERROR server.RepositoryServerWorker - Thread-2 (11): 97 failed
au.gov.naa.digipres.dpr.server.command.CommandException: CRITICAL: Failed to revert file movement to repository; manual resolution required!
at au.gov.naa.digipres.dpr.server.command.AddToRepositoryCommand.runCommand(AddToRepositoryCommand.java:520) ~[dprclient.jar:?]
at au.gov.naa.digipres.dpr.server.command.ServerCommand.runCommandWithEndMessage(ServerCommand.java:70) ~[dprclient.jar:?]
at au.gov.naa.digipres.dpr.server.RepositoryServerWorker.handleClient(RepositoryServerWorker.java:134) [dprclient.jar:?]
at au.gov.naa.digipres.dpr.server.RepositoryServerWorker.run(RepositoryServerWorker.java:91) [dprclient.jar:?]
Caused by: java.nio.file.NoSuchFileException: /repository/repository02/2017/04201233/pf_data/normalised
at sun.nio.fs.UnixException.translateToIOException(UnixException.java:86) ~[?:1.7.0_80]
at sun.nio.fs.UnixException.rethrowAsIOException(UnixException.java:102) ~[?:1.7.0_80]
at sun.nio.fs.UnixException.rethrowAsIOException(UnixException.java:107) ~[?:1.7.0_80]
at sun.nio.fs.UnixFileSystemProvider.implDelete(UnixFileSystemProvider.java:244) ~[?:1.7.0_80]
at sun.nio.fs.AbstractFileSystemProvider.delete(AbstractFileSystemProvider.java:103) ~[?:1.7.0_80]
at java.nio.file.Files.delete(Files.java:1079) ~[?:1.7.0_80]
at au.gov.naa.digipres.dpr.server.command.AddToRepositoryCommand.runCommand(AddToRepositoryCommand.java:488) ~[dprclient.jar:?]
... 3 more

The cause of the reversion turned out to be because the only active entry in the repository_location table (Repository02) was set to have a minimum space of 100GB which was more than the space available in the testing environment.

There are two issues here. The first is that the full details of the error are not returned to the client. The second is that the code to revert an unsuccessful transfer assumes that a normalised directory will be present when this is not always the case.

Discussion


Log in to post a comment.

MongoDB Logo MongoDB