Rename Failure Alarm
-
Getting a data source exception alarm from a CSV import in the form:
'DataSourceName':Rename Failure
whose status isNo RTN
The import works fine and the local files are renamed with the
imported_
prefix.Might be a bug or might just be me missing something; but I can't seem to trace the alarm to anything.
C
-
Hi cbyrne,
This is on the Data File Data Source. Is it possible the import directory has more than one file, and some do not have the imported_ prefix? Perhaps there are some hidden files in the folder? I would have expected there to be more information in the message, but it looks like there's a slight issue in the Data File Data Source/classes/i18n.properties file, specifically that the
{0}
for argument replacement is missing from thedatafile.event.renameFailed
translation key. Thanks for bringing this to our attention! You could modify that line to get the IOException message that triggered that event, with perhaps a Mango restart after modifying the i18n.properties file to be like,datafile.event.renameFailed=Rename Failure: {0}
But my first guess is that an
ls -al
may show a file owned by root or that the user running Mango does not have permission to (presuming Linux). -
Perfect, I'll do that now. Thanks @phildunlap! Also, while on the subject of the importer trying to import all files in a directory, would it be possible to apply a wildcard filter? Something simple like *.csv etc.
edit: Found the cause of the error. I had a script zipping up old imported_* files which placed them in a directory inside the import directory. The rename error was about the directory, duh. Thanks again for your help.
-
Currently no, but that is possibly a good suggestion. Currently the solution is to use individual directories for individual importers.
I'm surprised you report that to be the issue! I took a quick look at the code and it's checking if a listed file in the directory and shouldn't import in that case. I'll take a deeper look next week. You can have other files in the directory and have them be ignored, but only if they're named to begin with the imported prefix set on the data source, currently.