CIFS: cannot copy file: network name no longer available

[bart]

Hello,

When copying a bulk of files to the alfresco network share, I constantly get this message: cannot copy "filename": the specified network name is no longer available.

Sometimes: \\..\alfresco is not accessible. You might not have permission to use this network resource. Logon failure: unknown user name or bad password.

I manage to copy some files, but after some files have been copied, this message appears again and again. Quit annoying if you just have 30 days to evaluate alfresco …

thx

Bart

[bart]

Hello Gary,

Still having this issue …

I tried using an ftp client ie Filezilla. This works smoothly, only some caracters in the filename gave issues: trema, ampersand …

But when trying to open the space via the CFIS link, I still get an error message: "the path 'file:///\\share_a\… does not exist or is not a directory"

Thanks for your help guys,

Bart

[gary_spencer]

Hi,

There are fixes in teh 2.1 code to line up what the various protocols consider as illegal characters and what the repository considers are illegal.

Cheer

Gary

[bart]

Hello,

I'm using the community release 2.1.0 build 482

Cheer

[gary_spencer]

Hi,

Is this due there being an ampersand or other problem character in the file:// URL ?, it could be a URL encoding problem.

Cheers

Gary

[scottcarson]

Hey everyone,

Not only do I experience the same issue but we're up to 3.4d and the problem is STILL happening. In fact, it's been logged in a ticket:

http://issues.alfresco.com/jira/browse/ALF-561?page=com.atlassian.jira.plugin.system.issuetabpanels%...

The difference seems to evolve with the revisions of Alfresco, however. In the ticket (referencing 3.1 I believe), I was able to track specific errors. With 3.4, the problem is that the transfer runs, then croaks with the old "the specified network name is no longer available", BUT unlike the ticket the server doesn't come to a crashing halt. I can still go back to the CIFS share and browse and/or try the transfer again.

I've set up a test that I can reproduce every time:
- Connect to 3.4d alf server (kubuntu) via CIFS (win XP)
- Send/Copy 1 massive directory (1.37GB, containing 736 files, 93 directories) TO the alf server -> WORKS
- Get/Copy that same massive directory FROM the server -> FAILS
- Able to browse CIFS
- Restart alf server
- Get/Copy that same massive directory FROM the server -> FAILS
- Able to browse CIFS

So as you can see, it's an issue of direction (get/put). That said, it could also be  size or quantity (I'll test that soon). What's worse: my logs - catalina.out, alfresco.log, and webquickstart.log - all show *nothing*, as if everything is fine.

-Scott.

[scottcarson]

Update:

I was itching to try out one more test, so here's the latest:

- Connect to 3.4d alf server (kubuntu) via CIFS (win XP)
- Send/Copy 1 massive directory (1.37GB, containing 736 files, 93 directories) TO the alf server -> WORKS
- Get/Copy that same massive directory FROM the server -> FAILS
- Able to browse CIFS
- Restart alf server
- Get/Copy that same massive directory FROM the server -> FAILS
- Able to browse CIFS
- Get/Copy a smaller subdirectory of the massive directory (351MB, 249 files, 14 directories) FROM the server -> WORKS

Now, I can refine my findings to hypothesize that we have an issue of direction (get), and after more narrowing, size. I'm going to get my staff to see what the magic size number is (between 351Mb and 1.37GB) that causes the overload. Perhaps it's not size but rather the time it takes to transfer? Paolo may have been onto something with his suggestion of a buffer/timeout setting …

HTH,
-Scott.