User Name
Password
AppleNova Forums » Third-Party Products »

Anyone have experience with ZFS and/or TrueNAS or FreeNAS? (Or a Synology NAS?)


Register Members List Calendar Search FAQ Posting Guidelines
Anyone have experience with ZFS and/or TrueNAS or FreeNAS? (Or a Synology NAS?)
Page 2 of 6 Previous 1 [2] 3 4 5 6  Next Thread Tools
Brad
Selfish Heathen
 
Join Date: May 2004
Location: Zone of Pain
 
2022-07-31, 03:02

Correction to my last correction: TrueNAS CORE is based on FreeBSD, which has been their flagship product and was previously known as FreeNAS, but last year they also released a new Debian-based version called TrueNAS SCALE.

After playing around with both of them in VMs, I decided to go with the Debian-based TrueNAS SCALE because I'm a lot more comfortable with Linux than BSD.

Installing either of them was a major headache with my ancient motherboard, though, because Etcher, the drag-and-drop app that I usually use to burn ISOs to USB flash drives, for some reason failed to write the ISOs in a way that my old mobo would recognize as bootable. While troubleshooting, I copied ISOs for Pop!_OS and Fedora to the exact same USB flash drive, and they booted flawlessly. I also tried three different flash drives. So, in a moment of desperation, I resorted to using `dd` from the terminal to write directly to the /dev/ device… and that actually worked. Weird!

Once I fixed the flash drive boot issue, it was easy to get the rest up and running. Pretty much the same super basic installer as most modern Linux distros with very little user input or expertise required. Choose the boot drive. Define your root user password. Confirm detected network settings. Wait and reboot.

This old tower just has one 256GB SSD and two very old 1TB platter drives in it that may also be doomed to crash any day now, but that's perfectly fine for my experiment, and frankly I look forward to seeing how ZFS handles a drive failure. I created a mirrored pool of the two 1TB drives, defined some datasets in my pool, enabled SMB, copied a few video files over SMB from my Mac for testing, configured a Docker container to run Jellyfin, and boom! I have a NAS that's also a working media server! It even recognized my old NVIDIA GeForce GTX 960, passed the hardware into the container, and successfully used it for GPU-accelerated transcoding. Nice!

I'll grab some screenshots of the management UI to share later. Happy to answer any questions if you're curious.

The quality of this board depends on the quality of the posts. The only way to guarantee thoughtful, informative discussion is to write thoughtful, informative posts. AppleNova is not a real-time chat forum. You have time to compose messages and edit them before and after posting.
  quote
drewprops
Space Pirate
 
Join Date: May 2004
Location: Atlanta
 
2022-07-31, 03:57

Does the speed of your network connection matter? How fast is the Ethernet on that old Hackintosh?


...
  quote
Brad
Selfish Heathen
 
Join Date: May 2004
Location: Zone of Pain
 
2022-07-31, 14:42

I would guess that the speed of the old disks is more of a bottleneck in my case, but it's time for some science!

My former-Hackintosh-turned-NAS has 1 Gbps gigabit ethernet on its motherboard. Testing just raw data using iPerf3 (not touching any disks) over that connection:

Code:
❯ iperf3 -c 192.168.2.4 -p 9201 Connecting to host 192.168.2.4, port 9201 [ 5] local 192.168.2.1 port 53788 connected to 192.168.2.4 port 9201 [ ID] Interval Transfer Bitrate [ 5] 0.00-1.00 sec 112 MBytes 940 Mbits/sec [ 5] 1.00-2.00 sec 112 MBytes 935 Mbits/sec [ 5] 2.00-3.00 sec 112 MBytes 938 Mbits/sec [ 5] 3.00-4.00 sec 112 MBytes 940 Mbits/sec [ 5] 4.00-5.00 sec 110 MBytes 926 Mbits/sec [ 5] 5.00-6.00 sec 112 MBytes 943 Mbits/sec [ 5] 6.00-7.00 sec 111 MBytes 933 Mbits/sec [ 5] 7.00-8.00 sec 112 MBytes 936 Mbits/sec [ 5] 8.00-9.00 sec 112 MBytes 938 Mbits/sec [ 5] 9.00-10.00 sec 112 MBytes 942 Mbits/sec - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bitrate [ 5] 0.00-10.00 sec 1.09 GBytes 937 Mbits/sec sender [ 5] 0.00-10.00 sec 1.09 GBytes 937 Mbits/sec receiver iperf Done.
Meets maximum theoretical throughput expectations pretty closely.

Sending a large (~470 MB) file over rsync from my local system (which has a very fast NVME SSD) to the NAS:

Code:
❯ rsync -avPh ~/Downloads/OpenStep-Install-4.2-Dev.iso bsmith@192.168.2.4:~/ building file list ... 1 file to consider OpenStep-Install-4.2-Dev.iso 470.54M 100% 85.21MB/s 0:00:05 (xfer#1, to-check=0/1) sent 470.59M bytes received 42 bytes 85.56M bytes/sec
To compare, copying the same file locally, just to verify that my local SSD isn't slower:

Code:
❯ rsync -avPh ~/Downloads/OpenStep-Install-4.2-Dev.iso /tmp/ building file list ... 1 file to consider OpenStep-Install-4.2-Dev.iso 470.54M 100% 225.59MB/s 0:00:01 (xfer#1, to-check=0/1) sent 470.59M bytes received 42 bytes 188.24M bytes/sec total size is 470.54M speedup is 1.00
(yes, my local SSD is plenty fast by comparison)

And finally downloading the same file from the NAS to my local system:

Code:
❯ rm /tmp/OpenStep-Install-4.2-Dev.iso ❯ rsync -avPh bsmith@192.168.2.4:~/OpenStep-Install-4.2-Dev.iso /tmp/ receiving file list ... 1 file to consider OpenStep-Install-4.2-Dev.iso 470.54M 100% 111.49MB/s 0:00:04 (xfer#1, to-check=0/1) sent 38 bytes received 470.65M bytes 104.59M bytes/sec total size is 470.54M speedup is 1.00
Huh! Maybe my speculation about the old drives being a bottleneck was wrong because that's nearly saturating the 1Gbps ethernet connection. 111.49MB/s x 8 bits/byte = 891.92 Mbps, which is pretty dang close to the theoretical 1Gpbs maximum. The difference might just be the overhead cost of using rsync and ssh.

Since my two 1 TB drives are pooled as a mirror of each other, that means reading from the NAS should generally be faster than writing to it, but I was surprised to see such an immediately obvious jump from uploading (85.21MB/s) to downloading (111.49MB/s). I repeated these commands above a few times, deleting the destination files between each run, just to make sure it wasn't a fluke, and they all worked out to roughly the same numbers.

Regardless, the rest of the computers on my network are all on wifi, and that is a whole other bottleneck. Performing the earlier throughput test from a wireless device upstairs to the NAS wired into the router downstairs:

Code:
❯ iperf3 -c 192.168.1.120 -p 9201 Connecting to host 192.168.1.120, port 9201 [ 5] local 192.168.1.247 port 58521 connected to 192.168.1.120 port 9201 [ ID] Interval Transfer Bitrate [ 5] 0.00-1.00 sec 23.5 MBytes 197 Mbits/sec [ 5] 1.00-2.00 sec 23.1 MBytes 194 Mbits/sec [ 5] 2.00-3.00 sec 22.2 MBytes 186 Mbits/sec [ 5] 3.00-4.00 sec 22.5 MBytes 188 Mbits/sec [ 5] 4.00-5.00 sec 22.1 MBytes 185 Mbits/sec [ 5] 5.00-6.00 sec 21.5 MBytes 180 Mbits/sec [ 5] 6.00-7.00 sec 21.9 MBytes 183 Mbits/sec [ 5] 7.00-8.00 sec 22.4 MBytes 188 Mbits/sec [ 5] 8.00-9.00 sec 22.2 MBytes 187 Mbits/sec [ 5] 9.00-10.00 sec 22.4 MBytes 188 Mbits/sec - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bitrate [ 5] 0.00-10.00 sec 224 MBytes 188 Mbits/sec sender [ 5] 0.00-10.01 sec 224 MBytes 187 Mbits/sec receiver iperf Done.
Oof. 187 Mbits/sec is still pretty good for general web surfing and streaming, and it'll be fine for occasional file transfers and routine backups, but that's a far cry from the potential 937 Mbits/sec. This is why I still want to run ethernet cables around my house some day.

The quality of this board depends on the quality of the posts. The only way to guarantee thoughtful, informative discussion is to write thoughtful, informative posts. AppleNova is not a real-time chat forum. You have time to compose messages and edit them before and after posting.
  quote
Brad
Selfish Heathen
 
Join Date: May 2004
Location: Zone of Pain
 
2022-07-31, 19:00

Quote:
Originally Posted by Brad View Post
I'll grab some screenshots of the management UI to share later. Happy to answer any questions if you're curious.
…But then I realized that there's really not much interesting to show in screenshots. It's a pretty simple interface that looks and feels a lot like Cockpit, and it acts as a user-friendly wrapper for underlying commands with step-by-step forms to help novices.

Here's what my main dashboard looks like.



Other screens give you options to manage the disks themselves, various sharing services (SMB, NFS, etc.), users and credentials, system services (SSH, OpenVPN, etc), VMs (powered by KVM), "Apps" which are really Docker containers (powered by Kubernetes and Helm charts), and scheduled system maintenance commands. There's also a built-in CLI so you can run arbitrary commands through the web browser without SSH.

The quality of this board depends on the quality of the posts. The only way to guarantee thoughtful, informative discussion is to write thoughtful, informative posts. AppleNova is not a real-time chat forum. You have time to compose messages and edit them before and after posting.
  quote
turtle
Lord of the Rant.
Formerly turtle2472
 
Join Date: Mar 2005
Location: Upstate South Carolina
 
2022-08-01, 09:32

Thanks for keeping the updates on here for the TrueNAS. It is interesting to watch you work through the process.

It actually makes me want to test my speeds on my main NAS and see how well it is managing with the 10GbE NIC with the Mini that is 10GbE.

Louis L'Amour, “To make democracy work, we must be a nation of participants, not simply observers. One who does not vote has no right to complain.”
Visit our archived Minecraft world! | Maybe someday I'll proof read, until then deal with it.
  quote
Brad
Selfish Heathen
 
Join Date: May 2004
Location: Zone of Pain
 
2022-08-01, 10:51

In related news, thoroughly cleaning the dust out of and reapplying fresh thermal paste to a twelve-year-old CPU and stock Intel heatsink made a huge difference. Most of the case wasn't even all that dusty. A very fine powder in some areas, but no chunky dust bunnies.The case intake filters were a little dirty but not terrible. In the heatsink itself, though, I could see many of the gaps between fins were pretty clogged along the top where the fan connects.

In that screenshot I posted before cleaning, the CPU was reporting 53°C when mostly idle.

After the deep clean? It reports 38°C when mostly idle.

The quality of this board depends on the quality of the posts. The only way to guarantee thoughtful, informative discussion is to write thoughtful, informative posts. AppleNova is not a real-time chat forum. You have time to compose messages and edit them before and after posting.
  quote
Brad
Selfish Heathen
 
Join Date: May 2004
Location: Zone of Pain
 
2022-08-01, 10:53

Quote:
Originally Posted by turtle View Post
It actually makes me want to test my speeds on my main NAS and see how well it is managing with the 10GbE NIC with the Mini that is 10GbE.
I look forward to this!

10GbE sounds amazing, but I can hardly imagine ever coming anywhere close to saturating it. Which is probably a good thing.
  quote
turtle
Lord of the Rant.
Formerly turtle2472
 
Join Date: Mar 2005
Location: Upstate South Carolina
 
2022-08-01, 11:06

Nice!

I ran the network tests from the mini to the NAS (NAS running iPerf3 docker in server mode) and I'm not getting a fast as I had hoped... however this is still in production and at least two movies/shows were streaming at the time to various devices on my network. Not to mention there might have been some SFTP action going on from that server too. So, not a true test, more like real life speeds.
Code:
turtle2472@RackMini ~ % iperf3 -c 10.10.2.110 Connecting to host 10.10.2.110, port 5201 [ 5] local 10.10.2.53 port 63913 connected to 10.10.2.110 port 5201 [ ID] Interval Transfer Bitrate [ 5] 0.00-1.00 sec 920 MBytes 7.72 Gbits/sec [ 5] 1.00-2.00 sec 752 MBytes 6.31 Gbits/sec [ 5] 2.00-3.00 sec 782 MBytes 6.56 Gbits/sec [ 5] 3.00-4.00 sec 777 MBytes 6.52 Gbits/sec [ 5] 4.00-5.00 sec 770 MBytes 6.46 Gbits/sec [ 5] 5.00-6.00 sec 775 MBytes 6.50 Gbits/sec [ 5] 6.00-7.00 sec 775 MBytes 6.50 Gbits/sec [ 5] 7.00-8.00 sec 776 MBytes 6.51 Gbits/sec [ 5] 8.00-9.00 sec 762 MBytes 6.39 Gbits/sec [ 5] 9.00-10.00 sec 774 MBytes 6.49 Gbits/sec - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bitrate [ 5] 0.00-10.00 sec 7.68 GBytes 6.60 Gbits/sec sender [ 5] 0.00-10.00 sec 7.67 GBytes 6.59 Gbits/sec receiver iperf Done.
These are (Cat6) ethernet to SPF+ adapters into 10GbE ports on the same switch. The NAS is a DS1618+ with a PCI 10GbE adapter added aftermarket but default settings otherwise. "10000 Mbps, Full duplex, MTU 1500"

For drive speed tests, this is internal Mini SSD to internal Mini SSD:
Code:
$ rsync -avPh ~/Downloads/CentOS-Stream-9-latest-x86_64-dvd1.iso ~/ building file list ... 1 file to consider CentOS-Stream-9-latest-x86_64-dvd1.iso 8.43G 100% 256.85MB/s 0:00:31 (xfer#1, to-check=0/1) sent 8.43G bytes received 42 bytes 259.34M bytes/sec total size is 8.43G speedup is 1.00
This is from internal Mini SSD to NAS primary volume:
Code:
$ rsync -avPh ~/Downloads/CentOS-Stream-9-latest-x86_64-dvd1.iso /Volumes/home/ building file list ... 1 file to consider CentOS-Stream-9-latest-x86_64-dvd1.iso 8.43G 100% 166.15MB/s 0:00:48 (xfer#1, to-check=0/1) sent 8.43G bytes received 42 bytes 170.27M bytes/sec total size is 8.43G speedup is 1.00
For giggles, this is from TB3 connect SSD to NAS:
Code:
$ rsync -avPh /Volumes/T7/CentOS-Stream-9-latest-x86_64-dvd1.iso /Volumes/home/ building file list ... 1 file to consider CentOS-Stream-9-latest-x86_64-dvd1.iso 8.43G 100% 163.36MB/s 0:00:49 (xfer#1, to-check=0/1) sent 8.43G bytes received 42 bytes 170.27M bytes/sec total size is 8.43G speedup is 1.00
For extra giggles I pulled the file from the NAS to the ANMC server running on a host with a 1G NIC:
Code:
receiving incremental file list CentOS-Stream-9-latest-x86_64-dvd1.iso 8.43G 100% 79.06MB/s 0:01:41 (xfr#1, to-chk=0/1) sent 43 bytes received 8.43G bytes 76.98M bytes/sec total size is 8.43G speedup is 1.00
Less traffic on the NAS right now, closer to theoretical max:
Code:
[ ID] Interval Transfer Bitrate [ 5] 0.00-1.00 sec 848 MBytes 7.12 Gbits/sec [ 5] 1.00-2.00 sec 832 MBytes 6.98 Gbits/sec [ 5] 2.00-3.00 sec 1.03 GBytes 8.85 Gbits/sec [ 5] 3.00-4.00 sec 1.02 GBytes 8.73 Gbits/sec [ 5] 4.00-5.00 sec 1007 MBytes 8.45 Gbits/sec [ 5] 5.00-6.00 sec 1.04 GBytes 8.95 Gbits/sec [ 5] 6.00-7.00 sec 1.02 GBytes 8.75 Gbits/sec [ 5] 7.00-8.00 sec 920 MBytes 7.72 Gbits/sec [ 5] 8.00-9.00 sec 809 MBytes 6.79 Gbits/sec [ 5] 9.00-10.00 sec 759 MBytes 6.37 Gbits/sec - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bitrate [ 5] 0.00-10.00 sec 9.16 GBytes 7.87 Gbits/sec sender [ 5] 0.00-10.00 sec 9.15 GBytes 7.86 Gbits/sec receiver iperf Done.

Louis L'Amour, “To make democracy work, we must be a nation of participants, not simply observers. One who does not vote has no right to complain.”
Visit our archived Minecraft world! | Maybe someday I'll proof read, until then deal with it.

Last edited by turtle : 2022-08-01 at 21:10.
  quote
drewprops
Space Pirate
 
Join Date: May 2004
Location: Atlanta
 
2022-08-05, 20:50

I'm copying an external 4TB drive to a different external drive on my 2021 M1 MBP and still have 200,000 files left to copy after 23 hours.

Would it take this long on an NAS?

EDIT:
Nevermind, I read up on the app I'm using and it doesn't just copy, it checks the files and such and long transfer times like this are typical, so I'm over here talking kumquats while you all are talking apples - my bad
...

Last edited by drewprops : 2022-08-06 at 09:00.
  quote
Brad
Selfish Heathen
 
Join Date: May 2004
Location: Zone of Pain
 
2022-08-06, 19:13

Quote:
Originally Posted by drewprops View Post
I'm copying an external 4TB drive to a different external drive on my 2021 M1 MBP and still have 200,000 files left to copy after 23 hours.

Would it take this long on an NAS?

EDIT:
Nevermind, I read up on the app I'm using and it doesn't just copy, it checks the files and such and long transfer times like this are typical, so I'm over here talking kumquats while you all are talking apples - my bad
...
Actually, it's still a good question, and I suspect that the underlying mechanism your app is using for copying isn't too different from mine. I use the command-line tool "rsync" to do most of my remote file copying ("remote" being either to locally connected external drives or across a network), and it also does some pretty thorough sanity-checking along the way, estimates transfer rates, shows progress, and even copies all the weird Mac-specific metadata, attributes, and fork stuff. In general, copying lots of little files is much slower than copying fewer large files. Unless I know that I'll frequently need to dig into certain folders with lots of little files (like an archived email inbox), I'll zip the folder into one big file, and the time to zip it and transfer the zip is still usually faster than transferring the original folder of files. This is because systems have to stop to record where each file starts and stops and write metadata like created, modified, and access times for each along the way, and although that's usually only a very brief delay, those tiny delays will definitely add up to noticeable effects when you're churning through hundreds of thousands of files.

But it might be even worse for you…

Unfortunately, many "smaller" (1-8 TB) drives now use a new kind of technology called Shingled Magnetic Recording or "SMR" which can make performance truly awful in some conditions, including when you're doing lot of writing. Many 2.5" drives and some 3.5" drives, both internal and external, from pretty much all drive manufacturers have been found using this technology. I wouldn't be surprised if you got stuck with one; drive manufacturers do not advertise which drives do or don't use this technology. It's so bad that IT communities and companies like Synology and iXsystems (the TrueNAS company) have documented warnings that users and customers should not use particular drives because they basically can't work in a NAS due to the performance issues. There was even a class-action lawsuit in 2020, but I don't know if or how that ended. (I'd bet the manufacturers won because they always do.)

For now, it seems that only the larger 10+ TB drives consistently use the older Conventional Magnetic Recording or "CMR" technology which is much more performant across the board.

edit: If you're curious to know more about CMR vs SMR, here's a quick summary from Linus Tech Tips: https://www.youtube.com/watch?v=aztTf2gI55k Jump to 6:15 if you want to see some "holy shit" performance differences.

Speaking of which…!

Earlier this week I dropped a fat wad of cash on two 14TB external drives for my newborn NAS. Whenever I get new a drive, the first thing I do is connect it to one of my desktop computers and run a full two-pass secure erase of the whole thing. After forcing every block of every sector to be written, I'm reasonably confident that I didn't buy a lemon. It is a painfully slow process for these giant drives, though, even over USB3. The first one took almost 24 hours, and I haven't started the second one yet. Once they've both verified good, I'm going to shuck them to install in the NAS.

And I do plan to share the shucking process here. It's pretty neat, and it's something I never really considered doing until the last couple years. I've had many external drives over the decades, but without a robust long-term storage plan in place, which I think I have now for the first time ever, going from one external to the next was perfectly fine for my needs.

The quality of this board depends on the quality of the posts. The only way to guarantee thoughtful, informative discussion is to write thoughtful, informative posts. AppleNova is not a real-time chat forum. You have time to compose messages and edit them before and after posting.
  quote
drewprops
Space Pirate
 
Join Date: May 2004
Location: Atlanta
 
2022-08-06, 21:59

Quote:
Originally Posted by Brad View Post
Actually, it's still a good question,
THANK YOU for this reply, it was fascinating and informative.

It is incredibly helpful to understand that it's not a simple copy, but more along the lines of a Star Trek transporter, where it is pulling a full model of the structure into the transporter's buffer so that there's a mirror on the other end of the transport.

Somewhere along the line I have seen mention of the shingled drive format, and since these are indeed 2.5" drives I suppose the probability is high, especially since fewer people are probably using small externals and the stock could be old.

Watching this happen will be fun.

...

Steve Jobs ate my cat's watermelon.
Captain Drew on Twitter
  quote
turtle
Lord of the Rant.
Formerly turtle2472
 
Join Date: Mar 2005
Location: Upstate South Carolina
 
2022-08-08, 08:28

I started loving using docker compose to build containers on my NAS. There are some services running in an ubuntu VM right now that connect to my NAS via NFS and CIFS. If I can run those on the NAS in a container it relieves the network traffic for those files managed by the apps.

This is all well and good until the OOMs started happening.

I woke up this morning to my NAS being non-responsive as of ~0100. I had to hard bounce it. It recovered but I did leave, a java app is not a good candidate for the NAS even in a container.

Really, the issue is I have only 4GB RAM in the NAS and should have closer to 16 or even 32GB. I was swapping and all. Just not pretty.

HOWEVER!!!! The ability to run containers with a simple command and be able to use the GUI for basic restarts and stuff was great. I did move three of my most disk intense apps over to the NAS with stability so far so those will stay and the ubuntu server will remain active with the java based app and another that is just a memory hog.

Even still, not as good as it could be:
Code:
$ free -m total used free shared buff/cache available Mem: 3888 3170 138 22 580 415 Swap: 4383 2523 1860
Don't be cheap on your RAM for whatever NAS you get if you plan to do more than just basic network file sharing.

Louis L'Amour, “To make democracy work, we must be a nation of participants, not simply observers. One who does not vote has no right to complain.”
Visit our archived Minecraft world! | Maybe someday I'll proof read, until then deal with it.
  quote
turtle
Lord of the Rant.
Formerly turtle2472
 
Join Date: Mar 2005
Location: Upstate South Carolina
 
2022-08-08, 09:18

I just ordered 32GB of EEC RAM for my NAS. Not sure why I didn't do it sooner... but now is better than never.

Apparently Synology will nag you about not using their branded RAM, but it functions completely normally. I'll have to keep the original stuff just in case there is an issue I have to officially supported RAM for troubleshooting.

Louis L'Amour, “To make democracy work, we must be a nation of participants, not simply observers. One who does not vote has no right to complain.”
Visit our archived Minecraft world! | Maybe someday I'll proof read, until then deal with it.
  quote
Brad
Selfish Heathen
 
Join Date: May 2004
Location: Zone of Pain
 
2022-08-08, 11:22

That's a really good point about ECC RAM. In my research about all the wonders of ZFS, they do make it very clear that it's only as good as the hardware, and if a cosmic ray flips a bit inside RAM during a write operation, it's permanent and you have no way to recover unless you have a separate backup. Of course, it's no worse than a regular filesystem writing bad data, but the whole point of this is to be robust.

What I have in my old tower NAS is this: (purchased in 2015)
Corsair Vengeance Pro 16 GB (2 x 8 GB) DDR3-1866 CL9 Memory

What I'm looking at now is this:
A-Tech 16GB (2 x 8GB) DDR3 1600 MHz PC3-12800R ECC RDIMM 2Rx4 1.5V ECC Registered DIMM 240-Pin
(or maybe I'll splurge for 32 GB)

I thought ECC RAM was supposed to be a lot more expensive than “regular” non-ECC RAM. Have I misunderstood this for all these years? Or am I missing something else?

Also, every stick of RAM I've bought in recent memory has included an attached heat spreader (like that Corsair one). I see these A-Tech sticks don't have any. Any thoughts on that, T? Do you think it's worth grabbing a pair and attaching them? Or is the problem of keeping the RAM cool more a lot more overblown than it actually is, or something that only affects serious overclockers and hardcore gamers?
  quote
turtle
Lord of the Rant.
Formerly turtle2472
 
Join Date: Mar 2005
Location: Upstate South Carolina
 
2022-08-08, 11:33

If you are going to keep putting money into your old hardware be sure it is ECC compatible. Won't do you any good if the Mobo can't handle it.

For heat spreading... it is hit or miss. Most of the time I think it is more to do with looks rather than function. For all the sticks I've dealt with on the lower speed side there hasn't been heat sinks. Only once they start screaming.

Cost, yeah ECC normally is much more expensive. Here is a comparable though different brand for the ECC sticks you picked.

Edit: Here is a Crucial version of mine without ECC.. Almost as much for two DIMMs as I pay for a single with ECC. Laptop DIMMs seem to rob you so much more than desktop DIMMs.

Louis L'Amour, “To make democracy work, we must be a nation of participants, not simply observers. One who does not vote has no right to complain.”
Visit our archived Minecraft world! | Maybe someday I'll proof read, until then deal with it.

Last edited by turtle : 2022-08-08 at 12:42.
  quote
Ebby
Subdued and Medicated
 
Join Date: May 2004
Location: Over Yander
Send a message via AIM to Ebby  
2022-08-09, 02:00

Quote:
Originally Posted by turtle View Post
I just ordered 32GB of EEC RAM for my NAS. Not sure why I didn't do it sooner... but now is better than never.

Apparently Synology will nag you about not using their branded RAM, but it functions completely normally. I'll have to keep the original stuff just in case there is an issue I have to officially supported RAM for troubleshooting.
Word of caution, I did a big RAM upgrade on an old Synology and while it "detects" the full module, the CPU couldn't actually address that much. When I hit the RAM threshold, Plex would send garbled transcodes to users. I have no idea if that would happen to my data so I plopped the original RAM back in, scrubbed the drives, and everything went fine.

I don't know if your specific model has that limit though.

The Synology memory check didn't find anything wrong with either module. And yeah, I got the nag too. (And for a bunch of hard drives/m.2 SSD's)

^^ One more quality post from the desk of Ebby. ^^
SSBA | SmockBogger | SporkNET
  quote
turtle
Lord of the Rant.
Formerly turtle2472
 
Join Date: Mar 2005
Location: Upstate South Carolina
 
2022-08-09, 07:06

So I have to ask, does your version of NAS list supporting the RAM you put in it? I was reading about this before I ordered mine and my NAS specifically supports 32GB. Earlier stopped at 16GB IIRC. I'm REALLY hoping the nag is minimal and doesn't get worse over time... Companies tend to use scare tactics when there is no need to.

Louis L'Amour, “To make democracy work, we must be a nation of participants, not simply observers. One who does not vote has no right to complain.”
Visit our archived Minecraft world! | Maybe someday I'll proof read, until then deal with it.
  quote
Ebby
Subdued and Medicated
 
Join Date: May 2004
Location: Over Yander
Send a message via AIM to Ebby  
2022-08-10, 13:00

So one NAS has a INTEL Atom C3538 processor that can handle up to 256GB, but my second box has a Celeron J3355 that tops out at 8GB.

If Synology says it supports 32GB, you won't run into the addressing problem I did. I think you are safe there.

^^ One more quality post from the desk of Ebby. ^^
SSBA | SmockBogger | SporkNET
  quote
turtle
Lord of the Rant.
Formerly turtle2472
 
Join Date: Mar 2005
Location: Upstate South Carolina
 
2022-08-10, 13:44

Mine has the C3538 so I'm good to go with 32GB then! Oh, and it is out for delivery now! Tonight I'm upgrading!
  quote
turtle
Lord of the Rant.
Formerly turtle2472
 
Join Date: Mar 2005
Location: Upstate South Carolina
 
2022-08-10, 20:42

Holy crap this is taking forever. I upgraded the RAM in the NAS more than 30 minutes ago and I'm still waiting for it to boot. The power light is flashing but no ports open and DSM is unresponsive.


Off to DuckDuckGo and what do I find: this is normal for a DS1618+ after RAM upgrade.
Quote:
Startup delays are caused by memory checks. This is normal and will only occur once after each memory modification. Depending on the size of the memory installed, the check may take up to ten minutes. The Power LED will be blinking during this time.

Louis L'Amour, “To make democracy work, we must be a nation of participants, not simply observers. One who does not vote has no right to complain.”
Visit our archived Minecraft world! | Maybe someday I'll proof read, until then deal with it.
  quote
turtle
Lord of the Rant.
Formerly turtle2472
 
Join Date: Mar 2005
Location: Upstate South Carolina
 
2022-08-10, 21:49

..and I'm finally back up with 32GB of RAM.
Before and after:



I'm also not seeing an alert about the RAM not being Synology branded. Maybe it just hasn't nagged me yet, but I assumed I would see it in the info section or a notification on the sidebar if nothing else.

It took so long to come up because the RAM didn't seat right the first time so I had to wait it out to ensure I just wan't being impatient. Once the timer had gone long enough I pulled it with the intent to put the original RAM back in. The note reminded me to reseat and so I gave it a shot and it worked that time.

BTW, if you're curious, this is the NIC I have in my PCI slot. Plug and play.

Louis L'Amour, “To make democracy work, we must be a nation of participants, not simply observers. One who does not vote has no right to complain.”
Visit our archived Minecraft world! | Maybe someday I'll proof read, until then deal with it.
  quote
Ebby
Subdued and Medicated
 
Join Date: May 2004
Location: Over Yander
Send a message via AIM to Ebby  
2022-08-10, 22:27

Woot! looks good.

The Synology units do a memory test when the RAM changes so it was probably running though that for all 32GB. That would take a while.

EDIT: Duh, Ebby, read the whole post next time.

^^ One more quality post from the desk of Ebby. ^^
SSBA | SmockBogger | SporkNET
  quote
Brad
Selfish Heathen
 
Join Date: May 2004
Location: Zone of Pain
 
2022-08-14, 14:50

Dang… `badblocks` takes f o r e v e r to run on these large drives. I decided to verify my two new 14 TB drives "the linux way" just for good measure, even though I used macOS's Disk Utility to perform the 2-pass secure erase. Still in their USB enclosures, a ran `smartctl` short and then long. Once those succeeded, I ran `badblocks` destructively in a tmux session.

`badblocks` writes a specific pattern to the entire drive, reads the entire drive to verify that pattern, and then repeats with a different pattern. By default, the first pattern is `0xaa` or `10101010` in binary, and the second pattern is `0x55` or `01010101` in binary. The purpose of writing these opposite patterns is to verify that there are no bad blocks on the drive and that no bits get "stuck". The third pattern is `0xff` or `11111111` and the fourth is `0x00 or `00000000`, but I stopped after just the first two because writing and reading each pattern took almost two days, and that was enough to give me peace of mind and confidence to proceed.

So, after that week of churning, I can finally show off the "shucking" process that I mentioned earlier! Why shuck? These big external drives are often much cheaper than their bare internal counterparts even though they're basically the exact same hardware. Yes, this is counterintuitive. Yes, you might normally think an external drive with the added cost of the SATA-to-USB+power board, plastic shell, and external power supply would be more expensive. You would often be wrong. The economic and business reasons behind this are weird, and I'll try not to get into that here.



Here's the external drive after a nearly week-long battery of tests confirmed that it's 100% good.



I'm using celluloid guitar picks to non-destructively pry apart the case. You could try to use an old credit card or spudger or paint scraper, but guitar picks are way easier. I tried doing this without the guitar picks once before, and it was agony by comparison. Guitar picks are the perfect shape and hardness for this job. I got this set from Amazon for like $5, and they'll probably last a lifetime.



Slide the picks along the back edge of the case and gently rock them back and forth along these four positions. Inside the case are plastic clips at these locations that could break if you're not careful. You should feel a bump or a little resistance when you find a clip. The goal here is to slowly move the picks inside just enough to disengage these internal plastic clips.



Along the top and bottom near the front, use guitar picks to wiggle open a small gap until you can fit two credit cards stacked together. I suggest sandwiching two cards for prying here because the cards are larger (better for leverage) and using just one card is too flimsy. Alternatively, you could use a plastic spudger or screwdriver here too, but you might gouge the soft plastic.



With the clips successfully disengaged, The shell slides apart! Be gentle as you pull one half from the other. It's a tight fit, but remember that it's still just plastic. No metal clips or screws are holding this together.



Lifting the guts out of the shell.



Use a Phillips screwdriver to remove the single screw holding the board to the drive.



Pull out the light pipe. This is just a piece of clear plastic that transmits the drive's activity light from the board to the outside of the case.



Slide the SATA-to-USB board down off of the bare drive.



Remove the rubber shock-reduction bumpers.



Use a Torx screwdriver to remove the studs from the drive.



Done!!

Once you've done it once and know everything to do, it's really quite easy and takes maybe five minutes the second time. I'll be installing these in my makeshift NAS shortly, but I'm going to order that ECC RAM before I start storing anything mission-critical in there. It's getting close, though. I'm so excited to have this little project nearly complete.

The quality of this board depends on the quality of the posts. The only way to guarantee thoughtful, informative discussion is to write thoughtful, informative posts. AppleNova is not a real-time chat forum. You have time to compose messages and edit them before and after posting.
  quote
turtle
Lord of the Rant.
Formerly turtle2472
 
Join Date: Mar 2005
Location: Upstate South Carolina
 
2022-08-14, 19:08

Thanks for sharing that. WD is by far the easiest to get open. I've always gone with the cut up CC but might just invest in some picks. Worthy to have around since it is unlikely I'll buy plain drives with shucking being so much cheaper.

Louis L'Amour, “To make democracy work, we must be a nation of participants, not simply observers. One who does not vote has no right to complain.”
Visit our archived Minecraft world! | Maybe someday I'll proof read, until then deal with it.
  quote
Brad
Selfish Heathen
 
Join Date: May 2004
Location: Zone of Pain
 
2022-08-20, 13:38

Today is finally the day that I am installing drives into my makeshift NAS, but I just discovered something very odd when I opened her up to begin surgery. One of my SATA cables that's been in use for gods-only-know-how-many years appears to have… flecks of corrosion??

Some of the grime is clearly just old dust buildup, but that pale blueish greenish color looks vaguely like the kind of acidic corrosion you'd see on the terminals of an old leaky battery. Seeing this on a SATA connector, though, makes zero sense to me unless maybe the metal on the clip was somehow contaminated with something years ago and is just now showing the signs.





Has anyone ever seen connectors corrode like this before? Am I reading more into this than I should, and it's just greenish dust buildup? It's certainly not mold which would be my next (but even more distant) thought.

In any case, I'm tossing this one and replacing it with one of the many spares I have in storage, but I'm very curious to understand what this is.

The quality of this board depends on the quality of the posts. The only way to guarantee thoughtful, informative discussion is to write thoughtful, informative posts. AppleNova is not a real-time chat forum. You have time to compose messages and edit them before and after posting.
  quote
Brad
Selfish Heathen
 
Join Date: May 2004
Location: Zone of Pain
 
2022-08-20, 14:23

Following advice from another colleague who owns a couple Synology NASes, I simulated disaster by yanking one of the two old 1TB drives I had in this machine.

Fortunately, it did exactly what I hoped it would!

The TrueNAS dashboard and detail pages reports critical errors and warnings, but most importantly the mirror drive and everything using it are working fine while the pool is in "degraded" state!






The quality of this board depends on the quality of the posts. The only way to guarantee thoughtful, informative discussion is to write thoughtful, informative posts. AppleNova is not a real-time chat forum. You have time to compose messages and edit them before and after posting.
  quote
turtle
Lord of the Rant.
Formerly turtle2472
 
Join Date: Mar 2005
Location: Upstate South Carolina
 
2022-08-20, 21:43

I've never seen SATA cables like that while they were in the machine. After being trashed in a box maybe, but not in the machine.

Good to see you test it by pulling a drive. I need to do that again... but I likely won't.

Louis L'Amour, “To make democracy work, we must be a nation of participants, not simply observers. One who does not vote has no right to complain.”
Visit our archived Minecraft world! | Maybe someday I'll proof read, until then deal with it.
  quote
Brad
Selfish Heathen
 
Join Date: May 2004
Location: Zone of Pain
 
2022-08-21, 13:47

I discovered last night that my ancient CPU doesn't have hardware acceleration support for encryption, and this makes writing to encrypted datasets terribly slow.

Reading from an external drive and writing to the NAS with encryption? 30-45 MB/sec.
Reading from an external drive and writing to the NAS without encryption? 145-175 MB/sec.



Forget that! I just wiped everything to forego the encryption, and I started over. Previously I was looking at at least a couple of days to migrate everything into the NAS. On top of that, 3 of the 4 CPU cores were pegged at 100% to do the encryption in software, the temperature was hitting 70C, and the old fans were humming their hardest to keep up. I guess I deserve this for using 12-year-old hardware. After the wipe, I restarted my media files transfers just before going to bed, which previously projected over a day to complete… and it finished before I woke up.

I'm not terribly worried about keeping all my NAS data encrypted at rest, but my default M.O. is to encrypt all the things all the time. If we have a break-in, it's pretty unlikely that a thief would go for this over just about anything else expensive-looking in my house. ZFS defines "datasets" in a pool kind of like partitions in a disk, and you can define totally different settings (encryption, compression, deduplication, quotas, checksums, ACLs, etc) on each dataset or you can have some settings inherit from a parent dataset. It makes carving things up like this pretty flexible. So, now I'm thinking about segmenting my data differently. The big media files for Jellyfin? Definitely no encryption. Personal working documents (source code, Adobe projects, Minecraft backups, etc)? Probably no encryption. Private family documents, financial docs, and other keys? Absolutely getting encrypted.

Do the Synologys you guys have perform hardware-accelerated encryption on your data? I would imagine any CPU made in the last few years — even the cheap ones in these appliances — probably has the `aes` flag present. Would you take a look to confirm? `grep aes /proc/cpuinfo` should do it.

The quality of this board depends on the quality of the posts. The only way to guarantee thoughtful, informative discussion is to write thoughtful, informative posts. AppleNova is not a real-time chat forum. You have time to compose messages and edit them before and after posting.
  quote
turtle
Lord of the Rant.
Formerly turtle2472
 
Join Date: Mar 2005
Location: Upstate South Carolina
 
2022-08-22, 09:25

Both of my NASes support AES:
Code:
$ grep aes /proc/cpuinfo Features : fp asimd evtstrm aes pmull sha1 sha2 crc32 Features : fp asimd evtstrm aes pmull sha1 sha2 crc32 Features : fp asimd evtstrm aes pmull sha1 sha2 crc32 Features : fp asimd evtstrm aes pmull sha1 sha2 crc32
Code:
grep aes /proc/cpuinfo flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 sdbg cx16 xtpr pdcm sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave rdrand lahf_lm 3dnowprefetch epb intel_pt ssbd ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust smep erms mpx rdseed smap clflushopt sha_ni xsaveopt xsavec xgetbv1 dtherm arat pln pts md_clear arch_capabilities flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 sdbg cx16 xtpr pdcm sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave rdrand lahf_lm 3dnowprefetch epb intel_pt ssbd ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust smep erms mpx rdseed smap clflushopt sha_ni xsaveopt xsavec xgetbv1 dtherm arat pln pts md_clear arch_capabilities flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 sdbg cx16 xtpr pdcm sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave rdrand lahf_lm 3dnowprefetch epb intel_pt ssbd ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust smep erms mpx rdseed smap clflushopt sha_ni xsaveopt xsavec xgetbv1 dtherm arat pln pts md_clear arch_capabilities flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 sdbg cx16 xtpr pdcm sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave rdrand lahf_lm 3dnowprefetch epb intel_pt ssbd ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust smep erms mpx rdseed smap clflushopt sha_ni xsaveopt xsavec xgetbv1 dtherm arat pln pts md_clear arch_capabilities
The second one is the 1618+.
  quote
turtle
Lord of the Rant.
Formerly turtle2472
 
Join Date: Mar 2005
Location: Upstate South Carolina
 
2022-08-22, 21:46

Here is something you might appreciate:
Code:
sync; dd if=/dev/zero of=tempfile bs=1M count=1024; sync 1024+0 records in 1024+0 records out 1073741824 bytes (1.1 GB, 1.0 GiB) copied, 3.36027 s, 320 MB/s dd if=tempfile of=/dev/null bs=1M count=1024 1024+0 records in 1024+0 records out 1073741824 bytes (1.1 GB, 1.0 GiB) copied, 0.450154 s, 2.4 GB/s
Code:
sync; dd if=/dev/zero of=tempfile bs=1M count=1024; sync 1024+0 records in 1024+0 records out 1073741824 bytes (1.1 GB, 1.0 GiB) copied, 8.86652 s, 121 MB/s dd if=tempfile of=/dev/null bs=1M count=1024 1024+0 records in 1024+0 records out 1073741824 bytes (1.1 GB, 1.0 GiB) copied, 6.59869 s, 163 MB/s
The difference between the "+" and the "Value Series".

This is from a VM I just built to house my mariaDB server:
Code:
sync; dd if=/dev/zero of=tempfile bs=1M count=1024; sync 1024+0 records in 1024+0 records out 1073741824 bytes (1.1 GB, 1.0 GiB) copied, 0.411763 s, 2.6 GB/s dd if=tempfile of=/dev/null bs=1M count=1024 1024+0 records in 1024+0 records out 1073741824 bytes (1.1 GB, 1.0 GiB) copied, 0.145247 s, 7.4 GB/s
  quote
Posting Rules Navigation
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Page 2 of 6 Previous 1 [2] 3 4 5 6  Next

Post Reply

Forum Jump
Thread Tools
Similar Threads
Thread Thread Starter Forum Replies Last Post
Any one have any experience with an Eye-Fi SD card? turtle Third-Party Products 20 2010-04-29 13:06
Anyone got experience with ChipIn? chucker General Discussion 7 2006-08-21 00:26
Anyone have any experience with US Cellular? Partial Third-Party Products 0 2006-08-08 09:27
Anyone here have experience with... LoCash General Discussion 7 2004-08-18 18:05


« Previous Thread | Next Thread »

All times are GMT -5. The time now is 11:23.


Powered by vBulletin®
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Copyright ©2004 - 2024, AppleNova