Synametrics Technologies

Guys

Looking after some small business clients and migrating them from and old ESXI setup to ESXi 7 and using Veeam now for Backup

What we used to have and worked will for us (and them) using an older discontinued backup was as follows

1) Normal GFS backups of all VMs (5 - roughly 1.5TB roughly in size) - stored on site and portable drives taken offsite daily and then weeklies also on portable drives

We now want to modernize this and do the following

1) Veeam nightly backups - weekly full and daily incrementals

2) On the weekend peform a replication of all Virtual Machines to a Local USB drive - this means that we have full VMs ready to go with no restores etc in the event the VM Host crashed and we had to bring in a replacement server - performance would be slow to begin with - but we can then using Storage VMotion move the VMs back as bandiwdth permits. All VMs are stored with multiple files per disk (2GB option) 

3) Once the weekly replication to the USB drives has finalised we want to use Synchrify to perform off-site replication of the VM to our site where it will ultimately go onto SSD USB storage and we will keep multiple copies for rollback.

Can we schedule Syncrify so that it uses (say) 500Mb/s of bandwidth overnight on the wan link and then during the day (if it has not finished) can be throttled back to 200Mb/s ?

(They have 1GB links to the internet) 

Craig

Craig,

Yes, you can certainly do that. Did you have any questions?

One question: Will Veeam create a new file for every backup, or will it update the same files again? I ask because Syncrify uses rsync to minimize network traffic; in that case, no throttling is required.

Yes as we will be rotating the drives on a weekly basis on the remote site and using Veeam VM Replication jobs to copy the running VMs - then each (and all) of the files will be created new each time the job is run (and hence will need to be copied across the remote link)

I assume Rsync would when comparing the local and remote files see them as new - even though they are the same name and for the most part would be the same content 

Craig

Craig,

Syncrify will try to perform a block-match (use rsync) if the file names stay the same. Here is how it works in the background:

• Syncrify iterates through every available folder on the source. It processes one folder at a time.
• For a given folder, it will send a message to the server asking it to return a file of files along with their last modified date (LMD) and size.
• It will then perform a block-match for every file where either LMD or size does not match.
• The block-matching process results in a delta, representing the difference between the source and destination file.
• This delta is then sent to the server, which will merge the incoming delta with the older instance of the file.
• Eventually, both source and destination files will be identical. MD5 signatures are computed to ensure the files are identical

Therefore, even if you swap the drivers, it should work. Ensure you're NOT using Folder cache (https://web.synametrics.com/syncrifyfoldercache.htm). Using cache will not work with swapping drives.

Thats great and thanks for the detailed explanation - one last one before i put all of this together and get some testing done. These are quite large files and they are going across a VPN (yet to be determined what the production one will be - most probably WireGuard based) - is there a way to turn off the encryption on your SSH as it will drive the processors at each end over the top for extended periods whilst it is trying to encrypt all the data - whilst going over an encrypted link 

Craig

Craig,

Transfers in Syncrify are done using SSL. If you are already using VPN, you could use HTTP rather than HTTPS, which will stop in-transit encryption. When connecting from Syncrify Client to the server, ensure the URL starts with HTTP:// instead of HTTPS://