Importing Attachments into Jira Cloud via CSV Importer Not Working

[Hugh Flynn]

I am attempting to import issues from a very, very old copy of Redmine (as in not within the supported version range of the dedicated Redmine importer) and have gotten most everything I need to import correctly except for the attachments.

When the importer gets to the attachments, I can see it attempting connections (in the browser status bar) to the server where I have placed the files but each request shows an “An exception occurred dealing with attachment” error on the import results page.

A few points:

- I have the files on a basic web server running Apache.
- I can literally copy the URI from the Jira error page, paste it into a browser, and download the specified file so I know the files are publicly accessible with no authentication required.
- When I look at my web server’s log file, I see the requests from the importer are listed but the HTTP response code is a 301 redirect, which I’m assuming may be what’s breaking the import.

I’ve reviewed the Apache config and it is a largely default configuration (CentOS) and there are no redirects in either the .conf file or .htaccess file. Further, when the aforementioned files are downloaded via a browser, I see the expected 200 response code so there’s no redirect happening in that case.

I realize this is not an Apache support forum, but I’m hoping someone else may have run into this when trying to perform a CSV import. Are there any thoughts on why Apache would see requests from the importer differently than a browser, thus triggering the redirect?

[Brant Schroeder]

Hugh,

  Still sounds like the URL is not publicly accessible.  Just because you can access them from your workstation does not mean that you can access them from an external system. 

It could also be that the URL is a JIRA URL?  https://jira-instance/secure/attachment/...   If you log out of JIRA you will not be able to download the attachment.  The cloud application is not authenticated so it can not retrieve the attachment.   

[Hugh Flynn]

Thanks, Brant.

The URL is not a Jira URL. I'm migrating from a non-Jira system and the attachments are simply sitting on a web server. I'm taking this line from the the CSV importer documentation at face value:

"Note that URLs for attachments support the HTTP and HTTPS protocols and can be located anywhere that your Jira Cloud site can access."

I've even tried putting the files on a second server, just to eliminate that variable. I have accessed this second server via cURL using a completely separate external connection (a mobile hotspot) and I can successfully download the files.

Since the Jira importer doesn't provide much error detail, I can't really see what is happening when it attempts its connections for attachments.

[Brant Schroeder]

The links are public so anyone can download the files?  No authentication is required?  Have you tried testing a different URL like this one http://www.marketingmo.com/wp-content/uploads/2013/12/The-Strategic-Marketing-Process-eBook.pdf

[Hugh Flynn]

I just had a few test attachments available (much like the sample PDF in your last comment) so that I can get my issue export/import process worked out before importing the entire database of issues.

I was, however, able to spin up a cloud server, place my test attachments there, and have the Jira Cloud importer finally pull them in successfully. Clearly something with my prior web server configuration was interpreting the importer requests differently than those of a browser as I could never find an explanation for the 301 redirects.

Regardless, I've got a solution now. Thanks for your help.