Why use SnapRAID and direct connection

J_kel83 · 1. Juli 2019

I am new to OMV having just switched from a Windows 7 box with about 30TB of media. I have fully upgraded my server, now running dual Xeon in a 12 bay Supermicro with another 12bay Supermicro connected through my HBA. On my old system I had no form of RAID, I simply used drivepool to merge all my disks together and literally had a one for one swap in case any of the drives failed. I have continued that type of setup with my new OMV build, and I am trying to understand why I may want to use SnapRAID or if I would even get any benefit out of it. I understand if a drive fails I can toss in a blank one, click rebuild and poof a few hours (or days) later and all my data is magically back. But if I am already running a one for one swap it would be quicker to just toss in the “backup” drive and poof I am back up in running in a matter of minutes. Now I just have to get another blank disk and back up the back up. The only real advantage I could see is if I had not backed up my drives after ripping a few movies or uploading new photos. In either case my photos are still on the cloud and well the movies I could re rip if need be. So is there really any point to run SnapRAID in my case and if so what is it? Right now I do have it setup, but all I keep thinking is, I am giving up a drive and drive bay that could be used for more storage.

Second question, when I am ripping movies (ones I own I might add) I use another desktop (my OMV rig does not have an optical drive option), is it possible to directly connect those two via 10GBe cards to speed up the transfer process, while still keeping the normal Gig lan in place? I know it can be done in Windows, but am not that familiar with Linux (yet) nor OMV.

Thanks for your input
Jay

crashtest · 1. Juli 2019

First, for your pool of drives on Win7, I'm going to assume you didn't have 100% backup? With 100% backup and with an inventory or a list of the contents of a drive, restoration is a matter of coping data from backup.

And I'd have to ask a question, what do you mean by;

Zitat von J_kel83

I simply used drivepool to merge all my disks together and literally had a one for one swap in case any of the drives failed.

Did you copy the contents of a drive that was failing, to a new drive? Or did you restore from backup?
______________________________________________________________________________________

There's two ways hard drives drives fail; all at once (rare) or slowly (where data on the drive may be corrupted to varying degrees).
What did you do in either case? If a drive failed all at once, without backup, the content of the drive is gone. If a drive fails slowly, and backup is involved, the backup is slowly corrupted as well. Unfortunately, there's no fixing this kind of undetected corruption and silent "bit-rot" corruption once they occur.

Based on what you've said, with a media collection 30TB in size and with no active measure to detect bit-rot and other forms of corruption, it shouldn't be a surprise to find that you have media files with corruption, ranging from slight to significant. What might that mean? In a movie, there might be transient artifacts on screen or file corruption may be enough to stop the player. You may have seen both. Unfortunately, without some sort of active measure to detect and correct corruption, it may be noticed after it's too late. (When looking at a old movie, after 5 years.)

Along with file, folder and full drive restoration, detecting bit-rot and fixing it is something SNAPRAID does well, but that are limitations. SNAPRAID is designed for largely static storage, which is what you have, and it does restorations and detects/corrects bit-rot as of the last SYNC. This is it's one weakness. Since recovery is based on the state of your pool as of the last SYNC, files added since the SYNC are not protected and are not recoverable.

J_kel83 · 1. Juli 2019

Crashtest, thanks for the response. I had not really given much thought to bit-rot, or to that point even looked into what it was. I will definitely be doing more research before dismissing SnapRAID all together.

To answer your question, no I did not wait until a drive was failing to copy it, I would periodically sit down and back up each hard drive to another external, that would otherwise be stored in a disaster box. That way if an active drive did fail I simply pulled out that drives backup and popped it into the computer. The only data I had lost would be whatever had been ripped since the last backup. I did not know about SnapRAID at the time, and I was not fooling with regular raid.

Looks like I have some reading to do now though.

Jetzt mitmachen!