Help to set up a NAS?

The goblin

The goblin

<span class="poty">POTY Winner 2015</span></br>
Messages
4,407
Name
Marsha
Edit My Images
Yes
Hi

After months of shall I shan't I, I have finally bought a NAS. A Synology DS415+ with four 3TB WD Red drives.

On advice I have done the RAID 10 thingy and it took nearly a day doing stuff :thinking:

Now, as you can tell by my terminology I am not at all clued up on how these things run! All I know is I want to copy everything from my PC onto it. I want to access music and videos from it. More importantly I want to put all my photos on there. I have installed the photo station, video and iTunes server so hopefully my hubby and I can eventually have all our videos/ music accessible from the one place. The photo station I am thinking is best for putting all our favourite photos rather than my full catalogue!

But how is the best way to put my photo catalogue on it? As a backup from my PC. And is there a way that after I've added/ deleted/ edited photos on my PC that I can then sync the files?

I am fully aware that this is not a trustworthy backup system and if my house should burn down then I've lost the lot, but I would just like to get stuff on it, then I will buy another HDD to backup to and store away from the home.

Currently I back up my PC to my 2TB external HDD using Synctoy. Would I still need something like Synctoy with this NAS?

Any advice, tips, guidance very much appreciated as the Synology forum wasn't much help! I have a nerd friend that has helped fix my PC recently who could probably do it, but I really need to understand this beast myself.

Many thanks in advance.
 
strictly speaking a NAS isn't BACKUP - its redundancy...
Backup is generally something NOT located in the same area permanently.
however,

I only do a manual backup - could do a auto but lifes too short...lol

My photos are stored in a year, month, job basis.

I copy monthly (the previous months) folder to the backup drive from the NAS. Once its full, I'll just get another drive. (4tb)

All 2015 jobs are initially copied from cards to a portable 2.5' 2TB drive (in the field), then copied once home to the NAS (5 x 4TB) RAID 5, then monthly copied to the Backup drive...
 
NAS isn't BACKUP - its redundancy...
Click to expand...
Your getting confused with raid.
 
boolean said:
You could map a drive on your PC to the share on the NAS, then use something like allway sync to keep the files synced between the folder on your PC and the NAS.
Click to expand...

I have no idea how to do this!

Diving Pete said:
strictly speaking a NAS isn't BACKUP - its redundancy...
Backup is generally something NOT located in the same area permanently.
however,.
Click to expand...

Yup, I know it's not a true backup.

How do I get my files/ photos on it?

redsnappa said:
Your getting confused with raid.
Click to expand...

What is the difference between raid and redundancy?

Any websites out there that give guidance to setting a NAS up? I have Googled it to death!
 
What is the difference between raid and redundancy?
Click to expand...
The is no difference between RAID & redundancy. RAID is redundancy
RAID is an acronym which means Redundant Array of Inexpensive Disks.
 
Marsha

You should have DSM available to you which is Synology's desktop software. It is excellent.

Think about how you want to set up your folders on your NAS using the DSM dekstop. From memory I have something like the following as my shared folders then within each I have sub folders etc.

General
Books / PDFs
DVDs
Photos
Music
Time Machine
(probably more but I can't remember off the top of my head!)

I would use SynchToy to back up to the NAS Drive and then use the back up function in DSM to then back up to an external/removable drive which you can then take off site. The DSM function is easy to set up and is then a one click manual back up or you can set it up to complete on a regular frequency.


I use a Mac but my process is the same - back up via Time Machine for the Mac as a whole then then my important files including all of my photographs are backed up onto one of two external drives which I rotate, keeping one at work at all times.
 
From what I understand of Raid, with your 4x3Tb Raid 10 you have effectively one logical 3Tb disk?

I have a DS214Play - 2x3Tb disks in Raid 1 = effective 3Tb. I have a 3Tb disk attached to a USB port on the NAS and use the Synology back-up utility to back up the NAS to the USB external drive. I don't use SynchToy.

As to copying photos to the NAS, I just use the NAS 'Upload' command in File Station to copy my photos from my PC to the NAS. There's an option to overwrite or skip as the files are copied from the PC to the NAS. I skip.

So, all my photos are on the NAS, the USB external NAS back-up, my PC hard-drive, and an additional USB3 external hdd attached to my PC.

In terms of work-flow, I do whatever needs to be done to my photos on my PC, then simply copy those files off to the various storage devices. Viewing the photos on the NAS on other networked devices - PC's, TV's, etc - can be done via the PhotoStation software

The same principles are applied to music and video files.

FWIW, the Play variants of the Synology NAS are better for video as they have superior inbuilt processing capabilities.
 
Last edited:
Thanks Adrian, I have the latest DSM but it's not software I am familiar with so trying to learn how to use it before I mess it up and have to start again from scratch!

I have just popped back over to the Synology website and found this bit about backup and replication. The only trouble is they don't actually have any guidance for setting this thing up!
 
Last edited:
The goblin said:
Thanks Adrian, I have the latest DSM but it's not software I am familiar with so trying to learn how to use it before I mess it up and have to start again from scratch!

I have just popped back over to the Synology website and found this bit about backup and replication. The only trouble is they don't actually have any guidance for setting this thing up!
Click to expand...

Yes, it does take a little thinking through but it'll be worth it.

DSM is very powerful and can do lots of higher level stuff that you may ever need so you have to filter through that!!
 
With 4 x disks, I'd put it in RAID6, not RAID10. You get the same size, but with RAID6 ANY two disks can fail and the system will recover. With RAID10, if one disk goes you cannot lose its pair - you can only lose one of the other disks in the other RAID1 pairing. The other advantage is that with RAID6, you store parity information. This way, if one of the drives is failing to store data correctly, you know about it. With RAID10 you don't.
 
PS. We've just had this exact discussion in work about how to configure a compute server - we ended up with RAID6....
 
So what happens to a four-drive RAID6 in the unlikely event of two drives failing? I suspect you won't have a working array but I don't know what will have happened to the data.
 
So I have actually found a user manual for Synology DSM5.1, so I can at least try and read that and hope it all makes sense!

I have to say I am not an idiot, but a lot of this is like white noise to my brain! Give me an air traffic control tower and eight aircraft to speak to at the same time and I'm all over that, give me RAID talk and my brain cells melt!

So RAID 6 or 10 both give me the same capacity which is 6TB?

I don't understand the bit about drive failure!
arad85 said:
if one disk goes you cannot lose its pair
Click to expand...
If one disc goes can this one be replaced?
 
Yes, it can. RAID1 works by mirroring pairs of disk so you're protected if one of the pair fails. RAID10 takes two (in your case) pairs and splits the data across them for increased speed. If one disk fails, it's no problem. If two go and they're in different pairs then your data survives but, if two go and they're the halves of the same pair then you lose all your data. Modern drives are remarkably reliable and you have to be fairly unlucky to have one fail, let alone two.
I think four drives is too few for RAID6 as I don't know what happens in case of a drive failure - I suspect the remaining drives thrash horribly trying to recreate the two parity stripes, much like a three-drive RAID5 with a failure.
 
Last edited:
Snapsh0t said:
So what happens to a four-drive RAID6 in the unlikely event of two drives failing? I suspect you won't have a working array but I don't know what will have happened to the data.
Click to expand...
Yes you do have a working system. And you can rebuild it from the two drives. You also know when sectors are starting to fail, and can recover the data, unlike RAID1 where if one drive has a failing sector, the RAID1 array will return 2 bits of data which disagree. There is no way back from that.

Given RAID6, I don't really see the point of RAID1 past 2 disks.
 
Snapsh0t said:
I think four drives is too few for RAID6 as I don't know what happens in case of a drive failure - I suspect the remaining drives thrash horribly trying to recreate the two parity stripes, much like a three-drive RAID5 with a failure.
Click to expand...
Performance is degraded when the array is degraded, but that happens pretty infrequently - unless you decide that you're going to leave it in a degraded state for a long time (and to be honest, even in a degraded state, it's probably still fast enough to saturate gigabit Ethernet). As the OP hasn't done anything with the data, it would be possible to find out - configure as RAID6, fill it with lots of files and then pull a disk out and see what performance is like. You could then write to it in this state and then put the disk back in and see it rebuild the array.
 
The goblin said:
I don't understand the bit about drive failure!

arad85 said:
if one disk goes you cannot lose its pair
Click to expand...

If one disc goes can this one be replaced?
Click to expand...

Yes, but you are at increased risk of losing the data whilst you do this - especially if the drives are from the same batch and happen to be liable to failure (fcertain disks from manufacturers tend to be lower reliability, but you only find this out possibly years after they have been released). The problem is this. With RAID10, you create two RAID1 arrays and then stripe data across them. Say you have disk A, B, C and D paired AB and CD. If you lose disk B, you have a single point of failure - drive A. If you lose that whilst B is out, you lose all the data. With RAID 6 and you lose B, you can lose any one of A, C or D and still recover. The downside is, as has been said above, degradation of performance when one (or two) disks are out. Whilst this may be an issue for enterprise environments, it is not likely to be an issue for home use.
 
The goblin said:
So I have actually found a user manual for Synology DSM5.1, so I can at least try and read that and hope it all makes sense!
Click to expand...

Was this online? If so, can you share the link?

I spent ages setting up my Synology - and it ended up largely a hit and miss affair. I'm now sure what I did to get it working. Nor am I sure if it's working as well as it might be. So some destructions would be very handy.
 
Unless the OP is using No backup I'd run them as RAID 5 as that will allow one failure. If Backups are done frequently to another drive then this will allow the best bang for buck
 
Diving Pete said:
Unless the OP is using No backup I'd run them as RAID 5 as that will allow one failure. If Backups are done frequently to another drive then this will allow the best bang for buck
Click to expand...
As a storage engineer I wouldn't recommend RAID5 as per EMC's best practice for large NL-SAS (SATA Disks with SAS controller). RAID rebuilds with larger drives (1TB+) take a long time in which greater stress is being placed on the remaining disks (increasing probability of subsequent failure).
 
TheGrew said:
As a storage engineer I wouldn't recommend RAID5 as per EMC's best practice for large NL-SAS (SATA Disks with SAS controller). RAID rebuilds with larger drives (1TB+) take a long time in which greater stress is being placed on the remaining disks (increasing probability of subsequent failure).
Click to expand...
First I've heard of that... got any reputable proof that rebuilding RAID arrays causes higher failures (as it isn't my understanding ;))?
 
I am afraid I can't really provide proof apart from mine and my colleagues experience in dealing with enterprise class disk arrays. Regarding EMC best practice a quick Google should bring up documents but I am currently on my phone I can't grab these.
As a side thought another interesting study was done by I believe Google which also indicated that drives bought from the same manufacturing batch are statistically more likely to fail close to each other.
 
Oh the Grew is totally correct in this statment....

The larger the size drives the more likely the drives will fail especially if they are from the same batch / model...
This can be countered by drives with EEC (enterprise drives usually)...

BUT

its not applicable if countered by a regular backup !!!.... to a non-raid drive (& prefably kept offsite)

In context...

If a drive fails, add a new one... If the RAID then fails due to another drive failing then break the RAID, remove faulty drives & restore from backup drive /s.
Doesn't get much easier..
 
RAID Servers that are used by large companies are not really in the context of this POST !!!
The options that are open to large companies are far more than a couple of TOGS downloading their photos to a SMALL NAS server who may or not even know what a RAID is... - they may have got their mate to set it up....
 
Diving Pete said:
Oh the Grew is totally correct in this statment....

The larger the size drives the more likely the drives will fail especially if they are from the same batch / model...
This can be countered by drives with EEC (enterprise drives usually)...

BUT

its not applicable if countered by a regular backup !!!.... to a non-raid drive (& prefably kept offsite)

In context...

If a drive fails, add a new one... If the RAID then fails due to another drive failing then break the RAID, remove faulty drives & restore from backup drive /s.
Doesn't get much easier..
Click to expand...
Indeed enterprise drives have better MTBF ratings.
I also very much agree with the sentiment about backup (indeed the other half of my job).
Currently I backup to local NAS for fast RTO and recent RPO but then also take an offsite copy roughly once every couple of weeks that I keep at my desk draw at work. Cloud storage is also an option for this but my data quantities were too large for my upstream bandwidth.
 
& there is lots of data about RAID failures when using over 3Tb drives on the net...
This isn't a new issue.
 
I now operate on the basis of RAID - nice to have (& gives me a small amount of comfort)

BUT

not a substitute for BACKUP - prefabley offsite..

Too many people think because they have it on a RAID system that its safe - its the greatest con out there....
If your RAID controller dies - PSU spike, lightning etc that could easily kill your data.

If you NEED it then BACK IT UP.

Dont like the cloud, too slow to get your data, Hard drives are really cheap, don't make your life hard

My full time job is an IT engineer who also deals with the backup in our enterprise (650 people).
 
hollis_f said:
Was this online? If so, can you share the link?
Click to expand...
Yes, and Adrian kindly beat me to it :)

Diving Pete said:
Unless the OP is using No backup I'd run them as RAID 5 as that will allow one failure. If Backups are done frequently to another drive then this will allow the best bang for buck
Click to expand...
I do intend to back up. I do back up now to an external HDD, I just need to buy another to keep off site!

Diving Pete said:
If your RAID controller dies - PSU spike, lightning etc that could easily kill your data.
Click to expand...
This is a MASSIVE problem here in Cyprus. I have been within metre of three lightning strikes in 18 months! Our house got a direct hit in December, I was mid nap at the time but apparently sparks were seen coming out of the plug sockets! Thankfully the only thing fried was our router, our two week old £99 router :( Thankfully Asus just said send it back! No other item plugged in was fried and we try to unplug all delicate stuff during big storms that look like they're within striking distance!
 
TheGrew said:
As a side thought another interesting study was done by I believe Google which also indicated that drives bought from the same manufacturing batch are statistically more likely to fail close to each other.
Click to expand...
Yes they do. Here is a link to Google's study: research.google.com/archive/disk_failures.pdf Generally, temperasture was found to be far more significant than uasge... And they didn't find any correlation to usage vs failure - which was what you were claiming ;)
 
TheGrew said:
I am afraid I can't really provide proof apart from mine and my colleagues experience in dealing with enterprise class disk arrays.
Click to expand...
So what RAID level do you run?
 
Depends on the workload but for NL-SAS (the type of disk being used here) we would use RAID6 with the only exception being NetApp which uses its RAID-DP implementation which is based around RAID4
 
TheGrew said:
Depends on the workload but for NL-SAS (the type of disk being used here) we would use RAID6
Click to expand...
I was right to suggest that then ;)
 
No - in the context of the OP's post.

The OP has bought a 4 drive system.
RAID 6 is for want of a better word - crazy - in a 4 drive setup where the user has regular backups.

for those you dont know from Wiki
RAID 6 does not have a performance penalty for read operations, but it does have a performance penalty on write operations because of the overhead associated with parity calculations...

Losing half of your capacity just doesn't make any sense....
 
Diving Pete said:
for those you dont know from Wiki
RAID 6 does not have a performance penalty for read operations, but it does have a performance penalty on write operations because of the overhead associated with parity calculations...
Click to expand...
But if the performance of the system is high enough that this is invisible to you (as you will be limited by the network speed anyway), then it doesn't matter an iota whether there is a performance loss due to parity calcs. I run a software RAIDZ (RAID5 like) system with hot spare - it takes up about 5% of one core of an elderly AMD processor when flat out. I'm still writing to the array at 300MB/sec locally (I have 4 drives in the RAIDZ config), but this is limited to just over 100MB/s when accessing via gigabit Ethernet.

Diving Pete said:
Losing half of your capacity just doesn't make any sense....
Click to expand...
It depends on how belt and braces you want to be. I'm happy with RAIDZ as my data isn't that precious, plus the computer has a very good PSU and APC UPS. For this thread, I was comparing with the suggested RAID10 system where you also lose half of your disk space anyway. What I was saying is that it appears I was right to suggest RAID6 over RAID10.
 
Again,

in a Home system with only 4 drives - I wouldn't be installing either of those options....

What people do & the chioces they have in an Enterprise scenario are totally different that a small 4 drive NAS home setup.
 
I think the RAID level should be indicative of how valuable the data is to you combined with your ability to back it up. 6TB or even 9TB is an awful lot of data to back up to an ideally offsite location.
 
You must log in or register to reply here.
Back
Top