Revolving around the core of technology
Doing some testing with the syncrify server. Just doing local backups on the LAN (i7 on the client to NAS) the backups seem to take a long time to process:
https://i.imgur.com/ZvoZntC.png
These are the options used:
https://i.imgur.com/8F5g2BO.png
Is this what I should expect?
I've read http://web.synametrics.com/syncrifyandbw.htm and I wouldn't think it's the hashing even with the VSS enabled (this is Evo 850 SSD)
Evaluating for 150+ client install, 12TB of data. I'm trying to see if Syncrify can handle the data and the client load.
Thx,
David
David,
I have a few suggestions:
The amount of time for block matching is directly proportional to the size of the delta. Try selecting a few random files and see how big is the delta. You can do this by clicking the right mouse button on any file that has been modified and selecting Block Diff from the pop-up menu. If the delta is as big as the original file itself - meaning almost every block within the file is different, you're better off using Simple Copy instead of Delta Copy. Click here for details.
David,
I forgot to comment on VSS. VSS has no impact on how matching occurs. Once VSS is created, it becomes transparent, and the rsync algorithm does not know/care if it is being used. What did you mean by "Hashing"? I assume you meant "Block Matching", right?
Hashing...create a hash of the blocks of files to compare and determine what needs backup in a delta backup scenario. I assumed that was the only way you could match :)
Yes, I figured I could get some speed increases because of the LAN and disabling certain options...but I'm also wanting to test and see what kind of speed it will be when remote over WAN. I'll be having some clients with 500K+ files as well, and some with 2+TB of data with VM backups. If backups in those scenarios two extreme schenarios (once first backup is done) and backups run longer than 12hrs for an overnight, then I'm going to run into other problems later.
The restore process with versioning...is non-ideal :(
David
Ah, I get it now. To be more explicit, Syncrify uses MD5 hashes. These signatures are have nothing to do with VSS and will get created with or without VSS.
Check http://web.synametrics.com/syncrify-vhd-files.htm for some tips on backing up VHD files.
Syncrify works better on WAN than on LAN. I say that because the network is typically very fast on a LAN and therefore, the time spent on computing delta is often higher than transferring the entire file again without using rsync. Without rsync, Syncrify is no different than hundreds of other backup product out there.
It's the WAN where you need to minimize network traffic, and that is where Syncrify works better than any other backup product.
What do you mean by restoring versioned file is not ideal?
Well, I was going to try and keep this thread on topic but since you asked:
Just using the GUI of the Syncrify App is weird:
Then restoring:
Then Test Sync profile sync every minute:
....so I'm definitely running into more than a couple problems :(
David,
Profiles - Some backup programs call this Jobs. In short, a profile defines what gets backed up and when. Check http://web.synametrics.com/SyncrifyProfile.htm for details.
Adding folders/files - Two way to do that: either right click in the tree view on the left or drag-n-drop files from Windows Explorer. In my opinion, the second option is easier.
Log files - Log files are typically not for end-user. I guess you're used to seeing log files because you're technical. Therefore, this is no option in the GUI. An administrator can pull log files from any client machine from the server's web interface. Check http://web.synametrics.com/SyncrifyLogFiles.htm for details.
Opening Client from Systray - This is a good suggestion and I am going to forward this to our dev. team.
Restoring - Syncrify keeps multiple versions of file - not folders. This means file F1.txt could have 5 versions and F2.txt could have 1 version and both be in the same folder. Assume F1.txt changed for the past 5 days and you want to go back 2 days, you will end up getting version 3 of F1.txt and version 1 of F2.txt
You can restore previous versions of a file from the web interface. Right click on a desired file and select Previous Versions. Check http://web.synametrics.com/SyncrifyVersioningFeature.htm for details.
So by extrapolation of the logic "Restoring - Syncrify keeps multiple versions of file - not folders."
If you are backing up a shared drive for an office of people, with 500K files spread across 1000's of folders there no way for you to see the state (and restore said state) of any folder to an instance of time that a backup was run? A bulk restore operation to 4 days in the past will restore any file (deleted or otherwise) of every file that has ever existed in those folders regardless of deletion date?
...and then by the same logic, if you have a backed up cryptolocker style infection, how do you restore 500K files to the previous verions....or the version as it existed x days ago?
I apologize - I meant to say "Versioning" not "Restoring" keeps multiple versions of files.
It does not matter how many files you have 500K or millions. Each file will be independently versioned. The state of a folder is computed at the time of restore. Syncrify maintains the last modified date of each version. If you decide to go back five days, it figures out how far each file will have to get rollbacked.
I recommend you use versioning and see how Syncrify keeps the files on the destination (server's end). This will help you understand how it works.
Cryptolocker:
There are two variants: one that changes the files names and another that keeps the file names intact but changes the data within the file. You get protection from the first type by using the Delete Retention feature in Syncrify. Versioning protects the second category.