Any benefit to RAID 1?

t4keheart

Weaksauce
Joined
Sep 24, 2019
Messages
111
Hey guys,
I'm in the market for a storage solution, some sort of NAS. I'm torn between just buying a synology or a pc/server setup with some extra drives and doing it that way. I've had some low-end nas devices in the past, that I always ran in RAID 1, but now that I think or it, it doesn't make much sense.
My reasoning is that RAID 1 != backup. If a drive fails, you need to replace the drive (ideally with the same drive) and rebuild. Why would I not just take those same 2 identical drives, store files on one, and setup a backup to the other?
Wouldn't this be a more practical solution? Am I missing anything that a RAID mirror would provide that a incremental backup to a second drive wouldnt?
 
Raid 1 always has the same data on both drives at the same time. Doing backups every now and then requires more cpu cycles and there are times when the data is only on one drive.
 
RAID isn't a form of backup, in any of its iterations - RAID 1, 5, 6, 50, 60, whatever.

RAID serves three potential functions:
1- Resiliency against disk failure
2- Performance increase - more disks acting in concert = higher throughput
3- Increased capacity - combined disks to get a capacity greater than any single disk

You're primarily interested in function 1. While RAID isn't a form of backup, because any corruption from say a virus or user error would immediately affect both disks in the array, it *absolutely* provides a mitigation towards the downtime that might be experienced during a drive failure.

Example- you have two drives in your system, and a disk just died
Scenario 1- With RAID 1, you likely suffer zero downtime as your controller shifts the load to the second disk. You replace the failed disk as soon as is practical and rebuild the array.
Scenario 2- Without any form of RAID, you were periodically backing up the contents of your primary disk to a secondary disk. With the primary drive now dead, you must perform some form of restoration process. This can involve a fresh installation of Windows onto a new disk and then copying the data back, or perhaps starting the system using your backup software's bare-metal restore procedure to restore the backups to a new drive. During this process, your server is offline for some amount of time.

If the potential downtime of scenario 2 is more costly to you than the up-front cost of establishing a RAID array, then you should have a RAID array.

Ideally, you should simply have both if your data is critical. Run a RAID array for internal redundancy, and run backups as well - preferably keeping a copy of those backups offsite or offline.

This philosophy is useful in other areas. I have a client that once suffered an outage of their internet and phone service for ~12 hours. Due to the nature of the business, that client lost $70k during that 12 hour period. That event caused the client to reconsider the expense of having a redundant internet connection, which they had previously rejected as too expensive. At $600/month for another connection, that cost suddenly became worth it to them to avoid the pain of missing out on $70k of business. Cost vs risk :)
 
RAID isn't a form of backup, in any of its iterations - RAID 1, 5, 6, 50, 60, whatever.

RAID serves three potential functions:
1- Resiliency against disk failure
2- Performance increase - more disks acting in concert = higher throughput
3- Increased capacity - combined disks to get a capacity greater than any single disk

You're primarily interested in function 1. While RAID isn't a form of backup, because any corruption from say a virus or user error would immediately affect both disks in the array, it *absolutely* provides a mitigation towards the downtime that might be experienced during a drive failure.

Right, like i said in the op, i understand that no form of raid is equivalent to a backup. But for my purposes, I don't think any variation of RAID makes sense. The drives are for data storage only, no OS/booting. If I have one drive which backs up to the other once a day or whatever, if the primary fails I can simply change the secondary to the primary and continue backing up to another disk... highest risk is losing one days worth of data... which for my personal use isnt a big deal.

So would you agree then that, all enterprise/business level concerns aside, a simple 1 -1 backup between 2 drives makes the most sense to make sure my data doesn't disappear on me?
 
If it's data only, no installed applications or in-use databases or anything like that, then sure!
 
I'll reiterate though, I'd still recommend some kind of offsite backup. Backblaze is what I use, it's $5/month and sends all my personal data off to le internets. Worth it in case I get a cryptovirus that deletes all my shit or something.
 
If you had to pick only one then yes a backup is the safer solution. However unless you are dealing with huge capacity (expensive) drives I don't see a good reason to not run both.

One nice thing about a RAID 1 is that they are pretty much set and forget, particularly in a NAS. Your unit will start beeping at you if a drive goes. With backups you need to make sure you are monitoring them and ideally testing your ability to recover from them. Unfortunately recovery doesn't always go smooth.

I prefer both, a redundant drive to prevent me from having to recover and a backup in case things really go sideways.
 
For me, RAID 1 is just for reducing downtime. I have it on my home server using 2x 3TB drives that contain just documents. Everything else runs off of SSDs. Besides the cloud syncing with google drive and one drive, if one of the data drives goes down the odds are the 2nd one will still be running until I can get to swapping it out. Of course bad luck can screw both drives at the same time but what are the odds of that?

/knockonwood

#UPSallthethings
 
I'm torn between just buying a synology or a pc/server setup with some extra drives and doing it that way.
Unless your goal is to learn about these technologies, get the Synology. Specifically the Synology.

The drives are the easy part. You'll appreciate the reduction in size, heat, and potentially noise, while gaining an easy to manage solution that can be always on.
 
Unless your goal is to learn about these technologies, get the Synology. Specifically the Synology.

The drives are the easy part. You'll appreciate the reduction in size, heat, and potentially noise, while gaining an easy to manage solution that can be always on.

This x1000!
 
I second the synologies even my sister in law can use it and they use hardly any energy, one can deploy it in a loved one's closet and if it's not a frequently used closet they might not ever know where their backups go or where the LAN movies are other than accessible on their LAN.
 
OP, all you're suggesting is just a manual RAID 1 setup. You can't have a backup of a drive in the same the computer, it's not a backup at that point, it's just another copy of the data with the computer being the failure point for both your source drive and your "backup" drive. You may as well just use RAID 1 at that point. If you're going to make a backup of the data, it needs to be at least on a different system onsite and even better if you can have a copy offsite as well. If you can air-gap a copy, that's even better.

https://www.veeam.com/blog/how-to-follow-the-3-2-1-backup-rule-with-veeam-backup-replication.html
 
It is all about not loosing data.

In case of a fire, theft or amok hardware, external backup is the only solution.
I would call this the disaster case and for this you need a disaster backup.

But oldstyle backup is like old bread, always from yesterday and if you want versioning it becomes complicated unless you use ZFS replication as backup method with snaps. A Randomware attack that is detected after one or two weeks makes normal backup useless in many cases.

Mostly a data loss is due a failed disk, deleted files by accident, needed former versions or intentionally deleted, modified or encrypted files ex via Ransomware. The Raid 1 helps only against the failed disk and even then it can mean data corruption. This is due the write hole problem where all disks in a raid up updated sequentially. A failure during a write can mean a corrupted Raid, does not matter if Raid 1/5/6. Only new filesystems with Copy On Write like btrfs or ZFS can protect against (a hardware raid with BBU can help a little with old filesystems)

To really protect data, external backup is mandatory but in nearly all cases a good enough Raid level (with a mean time between dataloss within 10 years very very near to zero like 3way mirror or ZFS Z2/3) combined with data checksums to verify data and versioning via snaps like ZFS is offering is the way to go. In the last 12 years since I use ZFS I have not had to use my backups that I do via ZFS replication based on snaps you keep open files within or to run it very often per day on high load servers. All cases of data or storage problems could be solved via ZFS redundancy, checksum protection and snaps (use dozens of them, event thousands are no problem and they are readonly, Ransomware safe).
 
Back
Top