Revolving around the core of technology
I second this
I support this as well. There needs to be some sort of option for replication to DR over the web, possibly using rsync. Currently the methodology assumes you have bandwidth like in the states or that the DR will be on the same high speed network.
Paul,
One way is to use a Syncrify client to backup a Syncrify server. This approach leaves the server available for clients and you will be able to use rsync protocol.
It is important to understand why we recommend DR feature over a faster network. The rsync protocol reduces network traffic at the cost of local disk I/O and CPU. If Syncrify enabled the rsync protocol in DR feature, the server is going to do a lot of I/O after every backup. Disk is a probably the slowest resources on a computer. Imagine 20 processes trying to access the disk (which only has one head) at the same time. This will not only slow down the DR process but will reduce the overall scalability.
Imran
The recovery over slow connection is important to me, too.
I understand the topic from Imran.
I there a way to add a third option "mirror to slave", which is taken to a customizable time and interval?
Then I can set on e.g. "once per day at midnight". This is not a full disaster recovery, but a usable way for more little company or private installations.
The proposed solution does not convince me. The backup of the Repository as another profile in client to another syncrify server backs up all the .SYNVER files as normal files and slows down, because the offsite syncrify server checks each of these files.