[S3tools-commits] [Open Discussion] RE: s3cmd sync weird way of working

[SourceForge.net <no...@so...>]

The following forum message was posted by ludvigm at http://sourceforge.net/projects/s3tools/forums/forum/618865/topic/3748151:

Hi,

if you want to sync *from* S3 bucket to your local /var/www you must swap the arguments on B:
s3cmd sync --delete-removed -p s3://ebs-sync/var/ /var/www/

Otherwise s3cmd synchronises /var/www/newfile from A to S3, and then runs on B and sees that there is no /var/www/newfile - from s3cmd point of view it has been removed locally and because of --delete-removed it removes it from the bucket.

I guess you need one master server that synchronises TO S3 and a bunch of slave servers that synchronise FROM S3.

I hope I understood your problem right.