MadDiplomat

Thanks for the advice, I really appreciate it! I will stick to XFS, as it seems there is no point in rushing into early adoption of ZFS of array drives. Especially if speeds still need ironing out, as I will be getting rid of the last few 5400 RPM drives in my array.

Hello, I have an array of 7 drives with a single parity and XFS - encrypted FS. Does it make sense to break everything and use ZFS in the following situation? I've never used ZFS before, but I've heard endless of praise for it, so it must be the latest trend. Here's the situation: I want to upgrade to dual parity and replace about half of the drives with larger models. I could do this step by step, but it seems like it would take a really long time. So, I'm considering speedrunning it by backing up my entire array to another NAS, then destroying the array, installing the new drives all at once, creating a fresh configuration, and finally copying all the data back from the other NAS. Simultaneously, I'm thinking that I may as well format the "new" array's drives into ZFS with encryption (array drives can be encrypted if using ZFS, right?). I understand that since I will not use a ZFS pool, I won't get the best perks like bitrot healing, but supposedly, using ZFS for UnRAID's array drives still has certain benefits vs. XFS. If anything above is a bad idea, please stop me :).

No change with the latest UnRAID version. This plugin will put my system to sleep after the specified "extra delay" (in my case 99 minutes) and ignore other rules (array inactivity & network traffic), which were working as intended a few UnRAID versions ago. Is there any way to increase the delay beyond 99 minutes? It would make a useful workaround until this plugin is fixed. For some reason, the plugin will not let me input more than 99 minutes in the GUI (I'm not sure if this limit is intentional). Edit: I changed the "timeout" from 99 to 180 in /boot/config/plugins/dynamix.s3.sleep/dynamix.s3.sleep.cfg. This change is now reflected in the GUI, which previously did not allow me to enter a value greater than 99. I just hope it works and that I didn't break anything further. Edit 2: Well that did nothing. It says 180 minutes but it is still 99. Why is this even programmed like that is beyond me.

This plugin is broken now. Ever since I updated to UnRAID v. 6.11.1, it will go to sleep after the "delay" time, no matter what. I have increased the delay to 99 minutes and it just goes to sleep after 99 minutes. It even went to sleep while I was streaming a video from my server, ignoring all rules about array inactivity and traffic. It worked perfectly until I updated. I will now update to 6.11.3 and report if anything changes.

I'm happy to report that the effect of using encryption (xfs) on all array data disks and all cache pools has been negligible in my situation. Not only are the read/write speeds unchanged, the CPU utilization remains comparable to before. Note that I am not running any VMs, only a couple of dockers (Plex, downloaders ...). I only noticed one peculiarity, but I'm unsure if encryption had anything to do with it. To avoid shuffling the data and encrypting the drives one by one, I copied the entire array to another NAS, deleted all partitions and set a new configuration (essentially building the array anew) and then copied everything back. When I initiated the formatting of the 5 data drives to xfs-encrypted, the CPU usage spiked to 100% on all cores for about 3-5 minutes. Within this timeframe, the UnRAID web interface was completely unresponsive for about 2 minutes. However, this may simply be considered normal operation? The previous time I built an array from scratch and formatted all the drives was in 2011, so I really can't make a comparison

I just wasted 30 minutes on this. Today, 32GB flash drives are actually harder to find in stores than one would imagine. But even when I found one, the utility did not like it, even if I put it in a USB 2.0. port. I had to do the manual approach, but luckily the stick boots just fine now with UnRAID. If the app can't be updated, the legacy approach should be the default one presented on the web page.

Thanks for the info. I’m preparing to build my second unraid machine and once that’s done, I will encrypt the 2600K one. I’ll report back whether there was any impact on the speed, but it will probably take me a few weeks :)

I've been reading about the encryption feature but did not find much information about the expected performance hit. I'm worried that my aging 4-core system would be far too weak for it and that I should not even bother trying turning it on: B75 Pro3-M, i7-2600K, 4x4GB DDR3. I use 1 parity + 5 data disks without a cache drive. By using turbo writes, my write speed to the array is limited by the gigabit ethernet. I can already hit 50% CPU utilization if I, for example, use mc to copy lots of data and run some background docker apps. I'm thinking I don't have much headroom and seeing my speed drop to something like 50 MB/s with encryption would not be great at all. Thank you for sharing any experience or feedback.

Dear all, I would like to make a change: keep saving in-progress torrents to the cache-only user share, but move completed torrents to a user share in the array for seeding. My dilemma: can the mover be used for this purpose? Edit: Yes, it works.

I received the two WD Reds I ordered today and one of them is DOA. Tried it in three different computers - after power on the mechanism keeps spinning up and down and the system never detects the drive. I guess WD quality remains the same even after all these years. Makes me wonder if it was worth it at all spending the few extra coins on Reds. Greens still serve me well once the DOAs were replaced.

Thank you both for your input. Bjp, I guess you make a good point with the reliability. I'm not sure if I'm prepared to risk the potential hassle of DOAs in exchange for 10-20% lower price. Over the past several years I bought 6 WD green drives (not all of them were for unRAID) and two of them were DOA. Of course I received a replacement fairly fast and the new drive was then OK, but nevertheless the whole process is an annoyance. Edit: In the end, I opted for two WD Reds. Greens served my array for years without problems (except for those DOAs) and I expect these Reds to work even more reliably.

Hello all, I'm looking at a 3TB Verbatim HDD at one of the local shops - 3TB, 7200RPM, 64MB cache, product code 53166 (e.g. here). Here it is about 20 EUR cheaper than 3TB Seagate Barracuda or 3TB WD Green. However, I can't seem to find any useful information about these Verbatim internal drives on the Web. I presume Verbatim just resells these drives and the manufacturer is one of the big HDD companies? Does anyone have any idea what sort of drives these actually are and would the perform OK in an unRAID array? Thanks for any feedback!

I finished the upgrade. I can't believe how easy it was. I just swapped it into a new case, connected the HDDs to the ports in the same order as with the old motherboard and it worked out of the box like nothing changed. Even speedstep worked from the start. Awesome. Next month I think I'll start adding some much needed additional HDDs to the array.

Thanks! After considering the alternatives again (for my intended use of UnRAID, of course) I decided to order it, along with the Celeron G1620. I also threw in a Crucial 4GB DDR3 1600/1.5v memory stick for good measure. I'll report how everything worked out once I'll have set everything up and tested it in practice ... Based on your replies I don't expect any problems

Thank you. I think I might go with the ASRock B75 PRO3-M and Intel Celeron G1620. Combined they cost less that the GA-G1.Sniper A88X I was considering, LOL

Thank you both for your comments. I've been checking out different Intel Z87 boards with 8 or more SATA ports. Does UnRAID work well with the ASMEdia ASM1061 controller? Any problems or just plug and play? As far as I can see, the chipset provides 6 ports and then one or more ASM1061 controllers are added for additional ports, at least in the case of ASRock boards, and I'd like to know if that influences stability at all, etc. Thanks!

Hello all, My current (and 1st) UnRAID build served me completely error-free since 2011: - Asus AT3IONT-I - Patriot 2GB DDR3 1333MHz - 3xWD20EARS, 1xWD15EADS - Corsair Flash Voyager 2GB - Corsair TX650W (52A 12V) I am contemplating upgrading the motherboard and CPU, because I intend to finally upgrade my old clunky midi tower to Xpredator X3 big tower: - G1.Sniper A88X (AMD A88X Chipset) - AMD A6-5400K Is there anything special that I should be worried about or would you suggest a different (better) configuration around the same price range? I chose this motherboard because it has 8 SATA ports, the CPU doesn't really matter much to me, because I don't use CPU intensive tasks (so far, the Atom was sufficient for all my needs). Thanks for your replies!

When I was doing those tests Transmission was unloaded. Nothing was copying from or to the server. Which is why is find it so confusing. P.S. Oh for the love of sweet zombie Jesus. Transmission was the culprit. After setting file preallocation option to 2 (= full preallocation, default is 1 which is partial preallocation - this is what I was using so far), newly downloaded files to disk3 now transfer at ~400Mbps from disk3 to PC. Finally I can enjoy the awesomeness of unRAID again.

Great. What I am about to say will probably make no sense but if someone understands it please enlighten me. I use disk3 for Transmission. Copying files from unRAID that were created by it results in incredibly slow speeds. If I for example copy a downloaded 4GB file from disk3 it transfers at about 10MB/s. If I copy that file to disk1 or disk 2 and copy it from there to PC it transfers at about 50MB/s. If I copy that file to disk3 again (creating a duplicate) then it transfers to PC at about 55MB/s. ??? ??? ??? ??? ??? ???

My array: disk0 - 2.0TB WD Green - parity disk1 - 1.5TB WD Green disk2 - 2.0TB WD Green disk3 - 640GB WD Caviar Blue My problem: Copy test file, disk1 -> pc: ~360Mbps Copy test file, disk2 -> pc: ~360Mbps Copy test file, disk3 -> pc: ~120Mbps My 1st thought: the Blue is FUBAR. However, using hdparm -t /dev/sd?: /dev/sda: Timing buffered disk reads: 318 MB in 3.01 seconds = 105.76 MB/sec /dev/sdb: Timing buffered disk reads: 296 MB in 3.01 seconds = 98.24 MB/sec /dev/sdc: Timing buffered disk reads: 344 MB in 3.01 seconds = 114.32 MB/sec /dev/sdd: Timing buffered disk reads: 328 MB in 3.01 seconds = 109.12 MB/sec I am just so confused! I tried copying files off the array from 2 different machines and the same story repeats itself. My other HW is in the sig.

General airflow imho definitely contributes a lot. I use two 80mm fans at 5 volts to cool my 4 drives (5000-something RPM) in unRAID and their temperatures range between 33 to 38 °C when in use (depending on load and ambient temperature, although the latter doesn't change much due to aircon). The server does not have any other fans at all because the CPU is passively cooled. There was one occurrence when I forgot to plug in the power for the server's HDD fans. It was rebuilding parity at the time. When I checked the status about 2-3 hours later I saw that the disks reported 57°C. Now in my desktop computer I have two 7200 RPM drives (also stacked together); there is no additional cooling for them and yet during load (for example copying 1000GB of data from one drive to another) they will heat up to only about 40-42°C. The only explanation I have for this is airflow - that computer has a lot of fans - 4x140mm intake fans at the side (these really force in a LOT of air), 2x80mm exhaust at the top, 1x80mm exhaust at the back and also a 140mm CPU heatsink (transverse) fan that helps force the air out of the case.

Hi all, Things are finally resolved. As I mentioned the drive would cause the system to simply stop responding. After several attempts with WD's utility (which took me a few days) there was finally a result: I returned the drive along with these results and they gave me a replacement right away. The new drive has been successfully precleared today and I've just received unRAID plus key, so my 1st unRAID server will finally be operational very soon As some of you mentioned (and I agree with you), I should have just returned it right away. But since this was my first experience with unraid I thought that I'm doing something wrong when the drive stopped responding during preclear (the screen basically stopped updating, it was still at 0% after 6 hours).

Right. So the test never finished, aborted it after 2 hours. I tried the preclear script and nothing was happening with the pre-read, so I used -W to skip it. Step 2/10 (Copying zeros to remainder) has been stuck at 136MB (0%) for over an hour. Bah!

Or am I missing something here? I ordered two of these drives. One of them appears to be working fine. The other one caused my first ever unRAID (which I am in the process of building) to become extremely unresponsive. I hooked it up to my Windows machine and it took about 5 minutes to do a quick format. On the other drive it takes 5 seconds. Transfer rates are below 1 MB/s. And what's with the seek error rates? Still waiting for the results of quick 2-minute smart test from WD life guard app. It hasn't finished yet and it's been running for almost 20 minutes.

Alright, the stuff's all here, the week is over and I finally have some time to put it together, so thanks to everyone again for all your useful feedback, this community is great. I haven't put it all together completely yet, but I did a quick test and decided to stick with Corsair TX650W which I removed from my desktop (replaced it with HX1000W). Only AT3IONT-I and RAM connected (power consumption): Enermax EG465AX-VE(G): - System on (BIOS): 32W - System off, but plugged in: 3W Corsair TX650W: - System on (BIOS): 29W - System off, but plugged in: 1.2W I was quite surprised, because I expected the more powerful PSU to be less effective at lower draw. I'm very pleased that these Corsairs are worth the price.