Hi Michal

It's taken me sometime to dig into what went wrong. Still having difficulties.

This is what happened and where I'm at

 * downloaded files recursively using s3cmd put -r localdir/   s3://bucketname
 * a few days later...
 * ignored errors (see below) and so didn't realize directories were not copied 
(directory names created in top level as zero-length files)
 * reversed cmd (in moment of madness) and added zero-length directory name 
keys to S3
 * worked out using ls that both a directory key (e.g. 'folder/' and a file 
'folder') existed for each 'corrupted folder
 * manually deleted the 'folder' keys after which directory structure of 
key/values visible again via Cyberduck (my S3 browser)
 * continued to fail to recursively retrieve files


> Do your files still appear in 's3cmd ls --recursive' output?

Yes, all do.

> If they do there's a good chance you can retrieve them with "s3cmd get -r" or 
> "s3cmd get s3://.../absolute/path/to/the/file"

I can retrieve individual files but the recursive get no longer works.

I originally used put to download a copy of the bucket. A few days later I can 
no longer do a put or get to retrieve files recursively. Some do get 
transferred but most don't

With: s3cmd get -r <bucket> <localpath> 

I get :

mem...@dirtgirl:~/buckets$ s3cmd get -r --force s3://<bucketname>/ test3/ 
s3://<bucketname>/audio/assets -> test3/audio/assets  [1 of 1523]
 0 of 0     0% in    0s     0.00 B/s  done
ERROR: Skipping test3/audio/assets/190: Not a directory
ERROR: Skipping test3/audio/assets/190/original.wav: Not a directory
ERROR: Skipping test3/audio/assets/191: Not a directory
ERROR: Skipping test3/audio/assets/191/original.mp3: Not a directory

Any suggestions?


  .M.

-----Original Message-----
From: Michal Ludvig [mailto:mlud...@logix.net.nz] 
Sent: 03 December 2009 20:50
To: Michela Ledwidge
Cc: s3tools-general@lists.sourceforge.net
Subject: Re: [S3tools-general] Disaster recovery question

Hi Michela,

> Ive accidentally replaced an S3 key that was a folder <foldername> with
> a zero-length file <foldername>.
> 
> Can anyone advise how to repair?

I don't quite understand what have you done... S3 doesn't know anything
about "folders" - a folder is simply a filename with '/' in it. Thus a
concept of a standalone folder doesn't exist in S3.

Do your files still appear in 's3cmd ls --recursive' output? If they do
there's a good chance you can retrieve them with "s3cmd get -r" or
"s3cmd get s3://.../absolute/path/to/the/file"

Provide more details if you still struggle retrieving your data.

Michal




No virus found in this incoming message.
Checked by AVG - www.avg.com 
Version: 8.5.426 / Virus Database: 270.14.91/2541 - Release Date: 12/02/09 
19:43:00


------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing. 
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
S3tools-general mailing list
S3tools-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/s3tools-general

Reply via email to