In a previous thread on this forum, it states "iSCSI Volume Replication is synchronous and NAS Data Replication is synchronous and asynchronous. Block IO for iSCSI/NAS in latest version."
Perhaps I haven't understood exactly what's meant by this; I've create a NFS Share on two open-e DSS boxes (lets call it "test").
I've set box 1 to replicate this share to box 2, - under Maintenance -> Backup -> Data Replication -> we've left the "Replication whole files" check box unchecked (Log replication errors and Use ACL are still checked).
However when a large file (15GB) is being amended on box1, I can see it copying the whole 15GB file to box2.
I recommend to turn this feature off "Whole files" use when you have very small files 2-6k. It's recommended to turn it on, if speed of network is faster than local partition write speed.
I don't quite understand what you're saying; the wording on the option is "Replicate whole files" - the default is un-checked (ie off); when looking at the task it says "Replicate whole files: No".
The problem is that we are seeing the whole file being replicated, not just the blocks of the file that have changed.
This surely can't be correct? It certainly isn't doing what I would interperet the option to mean; and it pretty much makes trying to replicate our Virtual Machines to a remote site impossible if it's going to send their whole HD (file) each time.
You are correct - sorry responded to fast. When replication is started snapshot is made. After that this process compares changes between snapshot and destination then transfers the required data…If checksum is different then it will create new file with all your changes and replace with the old file. Data Replication needs snapshot and lot of system memory. Also try to dedicate a seperate NIC for replication.