What’s a NAS, and do I need one?

Topic

New Reply

ISSUE 18.41 • 2021-10-25 HARDWARE By Richard Hay If I were writing this to a group of aviators in the United States Navy, they would immediately respo
[See the full post at: What’s a NAS, and do I need one?]

1 user thanked author for this post.

SupremeLaW

Reply | Quote

Replies

Richard – thanks for nice article.  A few Qs:

1. Does the NAS back up in real time, or only on a schedule (like once per day at 3am)?
2. Is it smart enough to look only for file changes?
3. If “yes” to #2, does it act like Macrium Reflect incremental backups, or does it update individual files so that the NAS truly mirrors your original drive?
4. If I am using OneDrive for important documents, photos and other files, but they are also backed up on the NAS, are there conflicts and problems?
   1. Sub-Q:  If I use the default OneDrive settings so that the OneDrive docs live in the cloud and are not permanently stored on my PC, can the NAS still keep backup copies?
5. Can the NAS keep my huge .pst email file backed up and current?  We’re talking GBs in a single .pst file, and new emails might continue to come in at 3:10am while the NAS is backing up.

Thanks,

2 users thanked author for this post.

SupremeLaW, CraigS26

Reply | Quote

Hi Richard,

 

Instead of buying and installing your own NAS there are also cloud based NAS services which have a number of advantages: easier to install, your data is not lost when your house burns down and a better protection for ransomware. I use Pcloud, a swiss based service that offers a one time payment a 10 years  subscription. It is secure and, in the end less expensive then a home based system.

 

Arno

1 user thanked author for this post.

SupremeLaW

Reply | Quote

You may already have a ‘NAS’. If you have a higher end router there may be a USB connector on it. You can plug in a thumb drive or a USB hard drive. It’s very handy if you want to access your files when you are not at home. I don’t use this method since I have a mini-itx server running Linux. I can access my files using the VPN built into my router. Of course I backup my important stuff to a USB hard drive. Always have your files in at least two places.

2 users thanked author for this post.

wavy, SupremeLaW

Reply | Quote

>  Always have your files in at least two places.

I can’t stress enough how important it is to maintain at least two (2) identical copies of all important files.

When I did database programming many years ago, one of the first utilities we always wrote was a “file compare” program.  This program simply compared 2 files, bit-by-bit, to confirm whether or not the 2 sequences of zeroes and ones were exactly the same.

The first check that program would do is to compare the sizes of both files;  if the sizes were identical, then all binary digits were compared, bit-by-bit;  and, when finished comparing, that program would reply YES or NO i.e. identical or not identical.

We also honor a policy of downloading binary files from the Internet at least twice, e.g. device drivers in .zip format;  and, then we execute the Windows FC (File Compare) command to verify that both .zip files are identical.

The Internet is amazingly reliable, but occasionally a transmission is corrupted for one reason or another.

For example, in Command Prompt:

fc filename.1.zip filename.2.zip /b

 

FC can also compare 2 groups of files:

fc *.pdf N:\qnap\*.* /b

 

You can get FC help like this:

fc /?

replies with a list of options e.g.:

/B  Performs a binary comparison.

Reply | Quote

I’ve had two synology devices for many years. The 2-drive units are lunchbox size and easy to disconnect for emergencies. They are rated high availability.

1 user thanked author for this post.

SupremeLaW

Reply | Quote

Years ago I had a single 3TB drive NAS (with a spare 3TB drive cartridge) that was quite handy, but as it began to get long in the tooth, I started to think I would like more capability, more options.

I looked at NAS offerings and read reviews, etc. but finally decided I wanted my own version instead of someone else’s idea of what a NAS should be.  I built my DIY NAS in late 2015.  It’s a NAS/PLEX server, and does everything I envisioned when I was shopping for parts.

Last year I upgraded the drives, and it proved to be quite painless, with no downtime or loss of service, since all my SATA ports are hot-swappable.  It has proven to be very reliable and dependable.  Copying data and drive images is accomplished via Task Scheduler on my daily driver desktop,

Create a fresh drive image before making system changes/Windows updates, in case you need to start over!

We all have our own reasons for doing the things that we do. We don't all have to do the same things.

Computer Specs

2 users thanked author for this post.

aaron451, SupremeLaW

Reply | Quote

Can I add to the discussion on building your own NAS. If you are not a geek, and not building for fun, then a lot of the benefit comes from the NAS OS. I have had a Synology for 6 years now and there is no way I could have the functionality of their OS if I built my own.  It is a great deal better than the older one I replaced. It is horses for courses, depending on what you want, but for a tech literate amateur I would recommend getting a branded NAS.

Also see this thread for the practicalities:

https://www.askzwoody.com/forums/topic/what-a-difference-a-nas-makes/#post-2358903

Chris
Win 10 Pro x64 Group A

1 user thanked author for this post.

SupremeLaW

Reply | Quote

Many thanks, Richard:

VERY TIMELY!

If a DIY user has 2 PCs and a LAN switch, it’s pretty easy to simulate a simple NAS by sharing a folder on one PC, and mapping that folder to a network drive letter on the other PC.  Doing this is so easy, it’s an experience well worth having. You may need to fiddle a little with “permissions” on the shared folder.

Copying files to and from the shared folder can be done in Windows Command Prompt, using the XCOPY command, which works great over a LAN.  We enhanced that command with PUT and GET batch files that are designed to limit the scope of READ and WRITE operations to sub-folders in the source database of our large Internet website.

One of the really neat things about a modern high-performance NAS is their support for higher speed Ethernet (RJ-45) ports e.g. 2.5GbE, 5.0GbE and 10GbE.  And, if a user’s PC has one or more free USB 3.0+ ports, there are now plenty of USB-to-xGbE “dongles” that plug-and-play at speeds higher than the aging Gigabit Ethernet (“GE”) that’s been around for decades.

And, if a PC motherboard has empty PCIe expansion slots, PCIe NICs now support those higher speeds from several different vendors.

These higher speeds will be particularly useful for content creators who edit large video files that can be saved much faster to a NAS with high-speed connections, and then viewed by other PCs connected to the same NAS.

One last caution (in case anyone forgets): a NAS is just like an internal storage bay inside your PC chassis.  As such, it should be powered by a quality UPS/battery-backup unit, and that UPS should be cabled to a workstation that can monitor and record power events, like brown-outs and black-outs.  Contemplate the consequences if 2 x 12TB HDDs in RAID-0 are “bricked” by a lightning surge during a nasty winter storm!

Many thanks again to Richard and AskWoody for a very useful introduction!

 

 

Reply | Quote

I originally got an OmniNAS then upgraded to a Synology. They’re an excellent platform. I have 2 drives mirrored and have configured a media server. This way, I can browse photos, music and movies on my smart TV.

Quite easy to set up and good documentation.

To Glnz: NAS is network storage. Your questions are about backup software. You’d use other software that can backup to the network and choose a folder on the NAS as the storage location of the backup. If backup is your primary use, I’d mirror the NAS drives as mentioned in the article. Then you also have redundant backup.

A friend uses a NAS with hot-swap drives and alternates them for backups, moving one set off-site. Online backup is another redundancy option. I just backup key files that can’t be replaced that way.

2 users thanked author for this post.

Michael432, SupremeLaW

Reply | Quote

Depending on the application(s) in most frequent use, and depending a LOT on how much DRAM a user is willing to dedicate to it, a large ramdisk may be worth the time and effort required to install and tune it.

For example, a 64GB workstation user might dedicate half (32GB) to a ramdisk.

With UPS/battery backup units installed, the reliability of modern PC components is now good enough to depend on DRAM that is nearly fault-free.

A quality third-party ramdisk program we have used has a very useful option that updates the non-volatile image file on a user-defined interval.

It also maintains an internal cluster map that only updates clusters in the ramdisk that have changed — much like a virtual memory design that maps logical sectors to physical sectors.

As such, that image file and all of its folder contents can also be backed up on a user-defined schedule, just like any other files in the NTFS file system.

And, the most advanced ramdisk programs we have reviewed also support “mounting” and “dismounting” these ramdisk image files.

In this way, at day’s end (or sooner) there should be 3 identical copies of all files stored in such a ramdisk:

(1) ramdisk files and folders as stored in volatile DRAM

(2) files and folders as stored in the non-volatile ramdisk image file

(3) files and folders as stored in each copy of that ramdisk image file

Reply | Quote

The one place where we almost always install a ramdisk is the Firefox and Chrome browser caches.

So much of my day is spent accessing the Internet, e.g. Gmail, news, weather, the responsiveness of both browsers is very noticeable when the cache is resident in DRAM.

And, the contents of that cache can be lost completely withOUT any serious consequences;  both browsers simply renew cache contents, the next time it downloads any URL.

Lastly, shifting I/O loads to quality DRAM — typically sold with a lifetime factory warranty — offloads spinning HDD platters, and SSDs that need and enforce wear-leveling logic, because both are known to wear out much faster than DRAM.

Users of laptops with a spinning platter HDD should give serious consideration to a relatively small ramdisk e.g. 1GB, and add enough DRAM to make room for hosting a browser cache in that ramdisk.

Hard Knocks University, signing off 🙂

Reply | Quote

Yes!  Except for the latest DDR4, memory is cheap and it offers both performance improvements and reduced wear and tear on storage devices.  No surprise that my go-to Dell Precision T5810 has 32GB, soon to go to 64GB.  I have recently sold laptops tricked out with 64GB to power users.

1 user thanked author for this post.

SupremeLaW

Reply | Quote

>  Except for the latest DDR4, memory is cheap

Yes, and there will always be a “trickle-down” effect, as newer generations of DRAM come on-line.

According to one source, the DDR5 standard was published in July 2020, so it should be displacing DDR4 more and more with time.

A Dell Precision workstation with 64GB DDR4 would have been viewed as a “dream machine” 10 or more years ago!

Also, if MS has its way with Windows 11, there should be a lot more “refurbished” hardware showing up on-line at bargain prices.  I read that HP Enterprise recently persuaded MS to change their minimum h/w requirements for Win11.

Windows 10 should continue to work fine on “older” hardware when demoted to backup storage servers, where “older” means “Win11 no va (in Spanish)”.

Reply | Quote

Not mentioned in Richard’s article is the idea of building your own NAS.  There are perfectly functional free Linux-based NAS offerings, so it is not too hard to do.  But, at all costs, which may be considerable, AVOID used drives!   After all, a hard drive is the single weak point in a computer system that cannot be replaced, data and all.  If a motherboard goes out, so what?  It can always be replaced with another and a Linux-based NAS will start right up and go to work, after adjusting automatically for motherboard chipset, CPU and related drivers.

1 user thanked author for this post.

SupremeLaW

Reply | Quote

>  AVOID used drives!

Excellent advice! And, it doesn’t require too much time for all drives to become “used” 🙂

Which reminds me …

This may be a somewhat obscure point, but I still feel it’s worth mentioning.

In recent years, experts have honestly questioned the reliability of redundant RAID “re-build” logic.

Very briefly, if a given RAID level is chosen for its ability to automatically rebuild a failed array, it has happened that data corruption occurs DURING the RAID re-building process.  This forces a vicious circle that is never ending!

After reading about this kind of error, we’ve settled on RAID-0 for its speed and inherent wear-leveling;  and, with batch files we perform backups both within a single PC (using multiple drives and multiple drive letters), and across multiple PCs (using multiple remote drive letters).

And, we are performing a longitudinal hardware test by keeping XP PCs running, as long as possible, as backup storage servers:  the latter are only switched ON to update the backups, then switched OFF.  A PC that is SHUTDOWN cannot be infected with malware either.

Our objective seems to be feasible, namely, if a single HDD or SSD fails, we toss it without any regrets, replace it, and restore all of its original data from other drives and RAID arrays that have NOT failed.  Partition Wizard has a “copy partition” feature that works great.

For example, all drive images of any one C: system partition are archived on all discrete data partitions within that one PC e.g. images.001, images.002, images.003 are archived on E:, G:, I:, K: and so on.  D: is usually an ODD.  F:, H:, and J: store bootable clones of C: , so we can boot off any of those clones whenever needed — by changing the boot priority in the motherboard BIOS.

Graphically, any one PC’s storage subsystem looks like this (like Disk Management):

C: / E:

F: / G:

H: / I:

J: / K:

etc.

D:  optical disc drive

R:  ramdisk drive letter

X:  remote drive letter

Y:  remote drive letter

Z:  remote drive letter

Yes, there is that much redundancy in our LAN, which many observers might regard as excessive.

Nevertheless, some of our spinning platter HDDs are still running fine far beyond their factory warranties:  if/when they finally do fail, we’ll just toss them and replace them with brand new SSDs.

One might describe such a setup as a “multi-NAS LAN”.

Anyone who prefers Windows operating systems can usually find great bargains on REFURBISHED HP workstations at on-line retailers.  That hardware is robust, easily upgraded with PCIe expansion cards, and almost free when the retail cost of a new Windows OS is taken into account.

And, parts for REFURBISHED HP and Dell workstations are plentiful and easily located on the Internet.

Reply | Quote

Re Lightning. A NAS should always be unplugged when there is lightning in the area. No UPS can withstand a lightning strike.  There are other Defensive Computing suggestions for a NAS here

https://defensivecomputingchecklist.com/#nasdefense

Synology offers a free RADIUS server. I think QNAP does too. With it, you can upgrade your WiFi from WPA2 to WPA2 Enterprise (assuming your router supports it). Many IoT devices do not support WPA2 Enterprise, but it is a great option to have.

Get up to speed on router security at RouterSecurity.org and Defensive Computing at DefensiveComputingChecklist.com

1 user thanked author for this post.

SupremeLaW

Reply | Quote

>   No UPS can withstand a lightning strike.

All of our UPS units have survived occasional lightning storms.

I wasn’t implying a direct lighting hit e.g. directly on the UPS itself.

To me, that sounds more like a straw man, than a realistic expectation.

What I had in mind was a lightning storm centered several miles away (most are), which sends voltage spikes into the surrounding electrical grid.

And, occasionally a high-voltage spike may blow a transformer e.g. mounted on a power pole, adding more perturbations to the surrounding electrical grid.

Relatively low voltage spikes can and should be handled by a quality UPS.  For example, see the APC factory specs for that company’s warranties.

We also plug our UPS units into less expensive surge protectors, hoping that any excessive spikes will stop there before reaching each UPS.

And, to be extra careful, wall sockets can be replaced with GFI (ground fault interrupter) outlets that trip fast enough to prevent damages to downline surge protectors.

None of the incidents I had in mind are the kind that would MELT a UPS entirely!  🙂

Reply | Quote

I should also mention that all of our APC UPS units support user selection of settings like “sensitivity” and both the max and min voltage ranges that will trigger a switch to battery power.

These features are available in APC’s PowerChute Personal Edition software e.g. designed for SOHO users.

That software advises “high sensitivity” for “providing maximum protection against electrical noise”.  That’s the setting we always choose.

The more expensive APC units now output a pure sine wave, but I have no experience with any of those.

I also don’t have any experience with APC’s data center UPS units;  but, given the overall quality of their SOHO units, I would expect even more management options in their data center and server UPS units.

Hope this helps.

Reply | Quote

In the past, I have attempted to use a Raspberry Pi with an attached powered USB hub and some USB hard drives to create a custom NAS setup. The Raspberry Pi functions as a Samba server that can be accessed through my LAN by my Windows PCs. I am curious to know what you and others think of this idea.

Reply | Quote

I purchased a 4 bay NAS that said it supported NIC teaming so I could use 2GB of bandwidth for transfers. I have all 10G infrastructure. I would have preferred a 10G NAS but they are prohibitvely expensive in Australia (as is all technology). This “cheap” NAS cost me nearly $800.

It turns out only one connection could use one NIC at one time even when they were teamed so the best transfer you could get was 110MB/s (give or take).

My 2 bay $150 USB 3.1 Gen2 enclosure runs 220MB/s sustained transfers.

That Synology junk got returned real quick. Synology misrepresented it’s capabilities. It cannot perform as advertised. I also found that the 10G also did perform as described when I went and read reviews of the 10G products.

I will never waste money on NAS again.

I now have a 4 bay and a 2 bay USB-C 3.1 Gen 2 enclosure to go along with the 10 internal drives. No drive gets less than 200MB sustained transfer over the 10G network.

 

Reply | Quote

Hello All,

I’m fairly tech savvy but some of the NAS and Network Setting go beyond me. That said, I purchased several Buffalo NAS units and have them attached to my network. I originally had one which seemed to get wonky after a power fluctuation with multiple loss/regain of power. So, I bought a newer one and put them both on separate UPS units. Never could figure out how to restore the reading of the wonked unit. So, I did a reformat (I couldn’t afford to buy another big HD at that time) and used it for less important backups.

Anyway, starting last year, almost every time I boot up my computer Windows seems to lose some/all of the NAS “Mapped Drives”. I tried re-mapping them and several other suggested ‘fixes’ I found online, but they still drop out. So, I wrote a bat file which disconnects and reconnects the NAS units. With the upgrade from Win 10 to Win 11, I hoped this might go away but I still seem to lose the NAS drives.

Has anyone else been having this problem? Any solution you have found to cure it??

Thanks, Tom

Reply | Quote