IT Folks - A Question on NAS

[etrain16]

As mentioned, do a periodic backup of the NAS device and keep it elsewhere.

Fire, theft, or even a failure of the NAS device. Shat happens...

I'm still waiting for a single 10TB writable optical disc Or maybe a 10TB thumb drive...

 

[Vorpalis]

@etrain16:
RAID-1 just writes the same data to two or more disks.

Upsides:

• It only requires two disks, whereas other RAID types require more than two.
• It's computationally easy for the storage controller to do, so it doesn't require a RAID-specific controller. Even the the storage controller built-in to your motherboard can do it.
• It's fast to read from, because the controller can read the first chunk of data off disk 1 while simultaneously reading the second chunk off disk 2.
• Slightly more recoverable if more than one disk fails, assuming those disks don't contain redundant data for each other.

Downsides:

• No matter how many disks you have, you lose 1/2 of their capacity (n÷2).
• If you're using two disks, RAID-1 is as slow to write to as a single disk, because all data has to be written to both disks.
• It only works with even numbers of disks.

RAID-5 uses "parity data" for redundancy. Assuming we have three disks in RAID-5, what this means is that it takes the first chunk of data from each of the first two disks, and performs a computation called XOR or Exclusive-Or, then it writes the output to the third disk, in what's called the "parity stripe".

All About XOR: What XOR does is compares two bits and, if they're the same (0,0 or 1,1), the output is 0 (whether they're both 0s or 1s doesn't matter); if the two bits are different (0,1 or 1,0), the output is 1 (similarly, it doesn't matter which of the two bits is a 0 or a 1). For the visual learners:

Input Output
0 , 0 = 0
1 , 1 = 0
0 , 1 = 1
1 , 0 = 1​

How this gives you redundancy is that the output can be XOR'd with either input to give you the other input. Take our example again, only swap the symbols and run it backwards (the second input and the output become the inputs, and the first input becomes the output):

Input Output
0 = 0 , 0
1 = 1 , 0
0 = 1 , 1
1 = 0 , 1​

Running it backwards works the same if you use the first example's output and its first input to find the second input.​

Now back to RAID-5: So, with input 1, input 2, and the output each on separate disks, you can recover the original data if when a disk fails. Also, parity data for each disk is spread across all the other disks, so no matter which disk fails, its parity data survives on the remaining disks. (In our three disk example, parity data for disk 1 is spread across disks 2 and 3; parity data for disk 2 is on disks 1 and 3; and parity data for disk 3 is on disks 1 and 2).

Upsides:

• RAID-5 is more space-efficient than RAID-1. Instead of losing half of the total capacity to redundancy, you only lose the equivalent of one disk's capacity, regardless of how many disks are in the array (n-1). So this is a BIG deal on larger arrays. (If you have 10 x 1TB disks in RAID-1, your usable capacity is 5TB; but if you have those same disks in RAID-5, your usable capacity is 9TB).
• It can be nearly as fast to read from as RAID-1 (no parity computation is needed when reading).
• With a minimum of three disks, RAID-5 can be used on any number of disks, including odd numbers of disks.

Downsides:

• Because RAID-5 requires calculating parity data, it usually requires a purpose-built standalone controller. The chipsets on some higher-end workstation motherboards can do RAID-5, but it's generally slower than a standalone controller.
• Writes are slowed by having to calculate and then write parity data. The effect of this depends largely on how powerful the RAID controller is.
• You need a minimum of three disks to implement RAID-5.
• If more than one disks fails, you're SOL, but this is rare. (RAID-6 solves this by having duplicate parity data for each disk, stored on separate disks, but requires at least 4 disks to implement).

Synology Hybrid RAID I'm not as familiar with its inner workings, but my understanding is that it's similar to RAID-5. I do know that it's what we use on all our Synology NASes, and it's been robust and reliable.

Holler at me if you have questions.

 

[High Drag]

I bought and installed one of those for our fire department and I really liked it. Two things:
1) Get a UPS for it as suggested by another user (the NAS may autorecognize it via USB), and,
2) Update the NAS firmware after you get it.

Caveat...some online backup apps don't easily work with a NAS. Crashplan comes to mind.

 

[etrain16]

@etrain16:
RAID-1 just writes the same data to two or more disks.

Upsides:

• It only requires two disks, whereas other RAID types require more than two.
• It's computationally easy for the storage controller to do, so it doesn't require a RAID-specific controller. Even the the storage controller built-in to your motherboard can do it.
• It's fast to read from, because the controller can read the first chunk of data off disk 1 while simultaneously reading the second chunk off disk 2.
• Slightly more recoverable if more than one disk fails, assuming those disks don't contain redundant data for each other.

Downsides:

• No matter how many disks you have, you lose 1/2 of their capacity (n÷2).
• If you're using two disks, RAID-1 is as slow to write to as a single disk, because all data has to be written to both disks.
• It only works with even numbers of disks.

RAID-5 uses "parity data" for redundancy. Assuming we have three disks in RAID-5, what this means is that it takes the first chunk of data from each of the first two disks, and performs a computation called XOR or Exclusive-Or, then it writes the output to the third disk, in what's called the "parity stripe".

All About XOR: What XOR does is compares two bits and, if they're the same (0,0 or 1,1), the output is 0 (whether they're both 0s or 1s doesn't matter); if the two bits are different (0,1 or 1,0), the output is 1 (similarly, it doesn't matter which of the two bits is a 0 or a 1). For the visual learners:

Input Output
0 , 0 = 0
1 , 1 = 0
0 , 1 = 1
1 , 0 = 1​

How this gives you redundancy is that the output can be XOR'd with either input to give you the other input. Take our example again, only swap the symbols and run it backwards (the second input and the output become the inputs, and the first input becomes the output):

Input Output
0 = 0 , 0
1 = 1 , 0
0 = 1 , 1
1 = 0 , 1​

Running it backwards works the same if you use the first example's output and its first input to find the second input.​

Now back to RAID-5: So, with input 1, input 2, and the output each on separate disks, you can recover the original data if when a disk fails. Also, parity data for each disk is spread across all the other disks, so no matter which disk fails, its parity data survives on the remaining disks. (In our three disk example, parity data for disk 1 is spread across disks 2 and 3; parity data for disk 2 is on disks 1 and 3; and parity data for disk 3 is on disks 1 and 2).

Upsides:

• RAID-5 is more space-efficient than RAID-1. Instead of losing half of the total capacity to redundancy, you only lose the equivalent of one disk's capacity, regardless of how many disks are in the array (n-1). So this is a BIG deal on larger arrays. (If you have 10 x 1TB disks in RAID-1, your usable capacity is 5TB; but if you have those same disks in RAID-5, your usable capacity is 9TB).
• It can be nearly as fast to read from as RAID-1 (no parity computation is needed when reading).
• With a minimum of three disks, RAID-5 can be used on any number of disks, including odd numbers of disks.

Downsides:

• Because RAID-5 requires calculating parity data, it usually requires a purpose-built standalone controller. The chipsets on some higher-end workstation motherboards can do RAID-5, but it's generally slower than a standalone controller.
• Writes are slowed by having to calculate and then write parity data. The effect of this depends largely on how powerful the RAID controller is.
• You need a minimum of three disks to implement RAID-5.
• If more than one disks fails, you're SOL, but this is rare. (RAID-6 solves this by having duplicate parity data for each disk, stored on separate disks, but requires at least 4 disks to implement).

Synology Hybrid RAID I'm not as familiar with its inner workings, but my understanding is that it's similar to RAID-5. I do know that it's what we use on all our Synology NASes, and it's been robust and reliable.

Holler at me if you have questions.

Awesome info, thanks! I'll take some time to read that more in depth later.

 

[Member 5605]

@tarster

 

[ZigZagZeke]

For off-site storage of important files I use these guys:

Eskimo North (Seattle)
eskimo.com
1-800-246-6874
[email protected]

 

[tarster]

So reading through the list you lined out here, I have no prior knowledge with the Synology line but I have heard quite a bit of good stuff over the years.. About 2 years ago I bought a Asustor for a few reasons.. First, it was considerably cheaper for a 4bay nas then QNAP or Synology was, and second, I build computers on the side and have been working with computers for the last 13 years at home and 4 years for a company that is 24/7 up time with 99.999 uptime.. We have around 7k computers on the network about 2k printers.. But that doesnt equate to nas knowledge thats from my home use..

I ended up with the Asustor which I have been very very happy about, it was easy to setup and really easy to upgrade the disks down the road, when I first started I had 4x 500gb drives and then my wife filled them with photography stuff, so I upgraded them to 4x 2tb drives and thats what I am still spinning right now. I got the Asustor AS-304T which would allows me to run 4 drives and then also use the unit as a Media center if I want to.. I Currently have 3 cameras streaming motion detect files to it and then a 4th stream is a constant feed from one of the cameras.. Along with that. there are many apps you can install on the NAS to help give you more functions to it.. Such as.. Plex oh Plex is becoming a favorite friend of mine.. I stream all my tv shows and movies from it, Please.. look into Plex since you do alot of movie archiving.. you will archive your videos once.. and never put a DVD or CD into your machine again...

For the drives.. I would highly suggest the Western Digital Red line as they are designed for a NAS.

side note.. just because it has 2 network ports.. doesnt mean its going to actually help using both.. you always have to think of the bottle neck.. ok so you have two network ports on your nas.. then they both go into the router.. and the router has one port going into the cable modem.. oops.. yep didnt help. lol The two ports really only work if you have two separate networks running in your house, say one for a camera system and the second for the rest of your network.

The Asustor also has a number of Android and IOS apps that allow you to fully and remotely control your nas when your away..

UPS.. Please.. do not buy a nas with multiple disks.. and not buy a UPS.. thats shooting your self in the foot. I have a APC 1500 on my nas and cable modem and router, it will last at least an hour if not longer.. and with some notification emails setup, I can remote into my nas and shut it down if need be.

 

[DeanfromOregon]

Drobo is pretty awesome. Both of my IT guys use it for their home stuff. Drobo Data Storage Solutions, Network Attached Storage, Drobo Inc.

 

[PaulB47]

Secondly, I'd like it to be accessible when I'm away from home.

You might think twice about this. If you can get in from the outside, other people can too. It's a big security headache.

You might look at pfsense for your firewall. It's going to be a lot more difficult to hack than consumer-based routers and such, more on the level what banks would use. I've had pfsense for years now.

 

[tarster]

You might think twice about this. If you can get in from the outside, other people can too. It's a big security headache.

You might look at pfsense for your firewall. It's going to be a lot more difficult to hack than consumer-based routers and such, more on the level what banks would use. I've had pfsense for years now.

I know for my nas there is a white list and black list and after 3 failed logins it perm bans the IP.

 

[High Drag]

I don't think anybody hit the IP settings on the proposed NAS... so here's my take. Our Synology came with DHCP as the default. A couple of other choices you have are to set a static IP or leave the NAS set to DHCP and set your router to always give the NAS the same IP.

Maybe I'm doing something wrong, but when mapping a network drive to a drive letter in Windoze, I seem to have better luck (actually, better consistency over time) if dealing with a hard IP rather than the hostname or WTH the "friendly name" is. YMMV. My fave is to leave a device set to DHCP and use the router method.

 

[etrain16]

So reading through the list you lined out here, I have no prior knowledge with the Synology line but I have heard quite a bit of good stuff over the years.. About 2 years ago I bought a Asustor for a few reasons.. First, it was considerably cheaper for a 4bay nas then QNAP or Synology was, and second, I build computers on the side and have been working with computers for the last 13 years at home and 4 years for a company that is 24/7 up time with 99.999 uptime.. We have around 7k computers on the network about 2k printers.. But that doesnt equate to nas knowledge thats from my home use..

I ended up with the Asustor which I have been very very happy about, it was easy to setup and really easy to upgrade the disks down the road, when I first started I had 4x 500gb drives and then my wife filled them with photography stuff, so I upgraded them to 4x 2tb drives and thats what I am still spinning right now. I got the Asustor AS-304T which would allows me to run 4 drives and then also use the unit as a Media center if I want to.. I Currently have 3 cameras streaming motion detect files to it and then a 4th stream is a constant feed from one of the cameras.. Along with that. there are many apps you can install on the NAS to help give you more functions to it.. Such as.. Plex oh Plex is becoming a favorite friend of mine.. I stream all my tv shows and movies from it, Please.. look into Plex since you do alot of movie archiving.. you will archive your videos once.. and never put a DVD or CD into your machine again...

For the drives.. I would highly suggest the Western Digital Red line as they are designed for a NAS.

side note.. just because it has 2 network ports.. doesnt mean its going to actually help using both.. you always have to think of the bottle neck.. ok so you have two network ports on your nas.. then they both go into the router.. and the router has one port going into the cable modem.. oops.. yep didnt help. lol The two ports really only work if you have two separate networks running in your house, say one for a camera system and the second for the rest of your network.

The Asustor also has a number of Android and IOS apps that allow you to fully and remotely control your nas when your away..

UPS.. Please.. do not buy a nas with multiple disks.. and not buy a UPS.. thats shooting your self in the foot. I have a APC 1500 on my nas and cable modem and router, it will last at least an hour if not longer.. and with some notification emails setup, I can remote into my nas and shut it down if need be.

I appreciate that input and point of view. I'll keep that all in mind.

 

[etrain16]

Drobo is pretty awesome. Both of my IT guys use it for their home stuff. Drobo Data Storage Solutions, Network Attached Storage, Drobo Inc.

I will look into that, thank you.

 

[etrain16]

You might think twice about this. If you can get in from the outside, other people can too. It's a big security headache.

You might look at pfsense for your firewall. It's going to be a lot more difficult to hack than consumer-based routers and such, more on the level what banks would use. I've had pfsense for years now.

I know there are ways to make it very hard to get in, that's something I have to learn more about - I'll look into pfsense. I tend not to keep critical info on a drive like this, but I'd still like it protected.

 

[etrain16]

I don't think anybody hit the IP settings on the proposed NAS... so here's my take. Our Synology came with DHCP as the default. A couple of other choices you have are to set a static IP or leave the NAS set to DHCP and set your router to always give the NAS the same IP.

Maybe I'm doing something wrong, but when mapping a network drive to a drive letter in Windoze, I seem to have better luck (actually, better consistency over time) if dealing with a hard IP rather than the hostname or WTH the "friendly name" is. YMMV. My fave is to leave a device set to DHCP and use the router method.

Some systems I work with work better with static IP addresses, if that helps with NAS, I can certainly do that.