Volume problems NAS 314

Hi again

No problem. Happy to help.

Yes, the data recovery service is all remotely done. Only in very rare cases would NETGEAR ask for the unit to be sent in. There is no guarantee that they can fix it but it is worth having a discussing with them, I think. Even if they cannot recover for whatever reason then there are even more specialised services out there (a lot more $$) if the data is really really important. However, I think the outlook is decent enough for you here even though I have not seen the unit live. Might be advisable to power off the NAS until you get in contact with support to not add anymore strain onto those disks right now.

I am glad that you are considering keeping better up-to-date backups. I will tag Sandshark and StephenB as they are pretty good with advise on various backup strategies that they use.

Your (indirect) question about raid 5 is a good one. There are many debates from all corners of the Internet of this. Because the weakness of a raid 5 is during the re-sync of the raid when a disk need replacing. In that time period when the NAS is adding the new disk into your raid array, your NAS is vulnerable (in the most strenuous time as well) - as you found out. But then again, raid 5 is still a good overall solution in a 4 bay NAS because it trades redundancy for capacity well.

HDDs are typically very reliable and this issue you have is not a very common occurrence. A more common scenario is that people actually have multiple quite bad disks and then start replacing without realising that they are severely bringing their volume in jeopardy. Many also wait replacing disks until they are totally dead. I think that is a mistake and I my opinion any disk errors should be taken seriously. I personally replace disks with any sign of errors but some might say that is over the top.

I think a good option would be to look at the disk health of all disks prior to replacing a bad disk. You can download the logs and look at the disk_info.log

Here you will see all your disk's current health. Stats to be concerned about are:

ATA Error Count
Reallocated Sectors
Reallocation Events
Spin Retry Count
End-to-End Errors
Command Timeouts
Current Pending Sector Count
Uncorrectable Sector Count

I'd say if your disks exhibit any of these errors prior to accepting a new disk into the array - be sure to backup first!

---

tomatohead1 wrote:

Can you recommend a setup that is more reliable? Perhaps nothing works better than keeping this NAS and making sure back up is always in place. But if multiple disks can crash without warning, I would think there might be a more reliable solution.

 

---

To clarify - the vulnerability is in RAID itself, so any NAS from any vendor running single redundancy RAID has exactly the same issue.

RAID-6 can handle two failures, so it can tolerate having a second disk fail during the raid resync.  However, there are three costs:

• the write performance is slower
• the overall volume size is smaller
• expansion is more expensive, because all 4 disks need to be the same size

As far as the risks go - one factor here is that bad sectors aren't detected until you read or write to them.  In many cases (certainly on my own NAS) most of the files aren't accessed very often.  So it can take a long time to discover that there is a problem. 

OS-6 has several maintenance functions that can be run on a schedule - and two of them serve as good disk diagnostics.  One is the disk test, and the other is the btrfs scrub.  Personally I run each the maintenance tasks once every three months - and one reason is that I want to detect disk failures well before they get critical.

Hopchen wrote:

I am glad that you are considering keeping better up-to-date backups. I will tag @Sandsharkand @StephenB as they are pretty good with advise on various backup strategies that they use.

 

---

There are four general considerations that apply to all backup strategies:

One thing to consider is that sometimes the backup device itself can fail.  Sometimes I've needed to access a backup, and discovered that the backup couldn't be read.  A related issue is that the system might fail in the middle of a backup - and that can result in a corrupted backup. So my own approach is to have three copies of everything I care about (including the original).  I haven't lost data since I began that approach (long before I had a NAS).

The second thing is deciding "what you care about".  Obviously any personal content (photos, etc) is irreplaceable.  There's other content that could be recovered with effort (ripped movies or Music CDs for example).  And perhaps content that is fairly easily replaced.  But if you want to back up selectively, you need to be disciplined about keeping the backed up content separate from the unbacked up content.  Personally we're not going to be so careful.  So my own approach is to back up the entire NAS.  It's just easier.

The third thing is sorting out what threats to your data you will protect against.  Common threats are:

• Disk failures
• NAS failures
• User error (accidental deletion)
• Malware/virus - e.g. ransomware
• Physical threat - fire, flood, theft, nearby lightening strikes

Protecting against physical threats requires keeping at least some backups off-site.  Protection against ransomware also requires some isolation of the backup.

Finally, it's important to distinguish backup from two-way sync.  Two-way sync isn't a backup - in fact it increases the risk of data loss.  That's because deletions or modifications to any copy proprogate to all the others.  

Now on to specific options:

The cheapest backup method is to back up to USB disks.  Ideally you would use at least two rotating USB disks.  That protects against a failure of one of the USB drives.  You can protect against physical threats and ransomware by keeping one backup off-site - perhaps at a friend's house or in a safe deposit box.  You'd need to refresh that periodically, but perhaps not as often as the on-site backup.  If you use USB backup, make sure you format the drives so you can access the files from your PCs.  Don't use a file format that you can only access via the NAS.  Also, I suggest making a full backup periodically - don't just do incremental for years.  The full backup can uncover failures in the USB drive that otherwise would be missed.  The main disadvantage to USB backup is that you end up handling the disks fairly frequently - so it's not fully automated.

Another method that can be cost effective is Cloud Backup.  You do need to be careful here, as many cloud storage options are two-way sync, and not true backup.  One advantage is that Cloud Backup is off-site, so it addresses physical vulnerability.  Many cloud services also provide reasonable protection from ransomware.  There are two disadvantages - it is pretty slow (both backing up and recovering), and it might not be fully reliable. 

The most expensive (but most convenient) method is to back up the data to another NAS.  This can be fully automated, and you can expand the backup volume as you expand the main NAS.  You can get some protection from ransomware if you only enable rsync on the backup NAS (since the PCs won't be able to directly infect it).  If a friend cooperates, you can even place a backup NAS off-site.

What I personally do:

My main NAS is an RN526x.  I have snapshots enabled with 3 months retention, which gives me reasonable protection against user error.  I back this up using daily backup jobs to two other local NAS - a Pro-6, and an RN524x.  That gives me three local copies of everything.  All the NAS are protected by UPS, which eliminates the possibility of losing data due to an unclean shutdown and offers some protection from power surges.  The RN524x only has rsync enabled, and is on a power schedule - so it is off when it isn't making backups.  That provides some ransomware protection. I'll likely do the same with the Pro-6, I just haven't gotten to it yet.

In addition, I run CrashPlan (Cloud) Backup on a PC that has the RN526x data volume mapped to a drive letter.  I'm only backing up the NAS, so I remain within the terms of service.  This gives me disaster recovery from the physical threats.  I've been using Crashplan for some years now - it's been generally reliable, and I find it affordable ($10 per month).  However, I still see some risk in relying only on a cloud backup service.

That said, the addition of CrashPlan actually gives me four copies of everything - so at some point I could drop one of the ReadyNAS backups.  However, there's no need to consider that until one of the NAS fails.

Using two backup NAS is a fairly expensive, but in my case it developed over time.  Initially I was backing up my NAS to dedicated hard drives in two desktop PCs - similar to using USB drives. As I outgrew NAS, I'd repurpose them as backup NAS - backing up at least some of the shares of the new main NAS to the older one.  For a while I was distributing share backups across multiple older NAS and PC hard drives, as I didn't enough storage in any single device to back up the full NAS.  Eventually I ended up with two backup NAS that are large enough to back up the main NAS.  I guess my main point here is that you should grow your backup solution as you grow your storage.