Jump to content

Talonius

Members
  • Content Count

    5
  • Joined

  • Last visited

Community Reputation

0 Neutral

About Talonius

  • Rank
    Newbie
  1. First, what are the implications of switching a motherboard, CPU, and RAM for UNRAID? The installation is tied to the key which is tied to the Flash device which boots, right? Not to a particular hardware signature? (UNRAID Plus, if it matters.) If I switch motherboards all of my drives will have new identifiers. Is it as easy as placing them all in a new array? Should I ensure that each drive is placed back in the same position within the array? (Probably so, just to be safe.) I'll have to recalculate parity after the installation, I'm sure. Are there implications I'm not aware of or thinking of here? I'd like to do this as painlessly as possible - that means not losing 60TB of data. (And I have no way to back it up... only about 36TB of other storage space available and that's impaired until I get a replacement drive.) Second, the actual switch. Right now I have an Asus M5A78L-M/USB3 motherboard (AM3+), 32GB of RAM, and an AMD FX-6300 Hex Core at 3.5Ghz. My server is switching off at random intervals, but noticeably during heavy load. I believe my CPU is overheating and the shutdown is the "oh shit, save the system" operation. The cooler that's currently on the CPU is vendor trash and I can replace it, however, I believe a new CPU and motherboard might enhance the usability of my system. Right now with parity check, a Sonarr docker, a Radarr docker, a Deluge docker, a NzbGet docker, and a Jackett docker my CPU usage sits at 100% (via UNRAID Main screen; via HTOP it's about 75% with a lot of kernel [IOWAIT] time.) My temperature sits at 79C/174F; the hivemind of Google says I should keep it below 77C/170F. My performance is compromised with these dockers - Deluge can kill a computer on its own most of the time - but if I go to remove a file with SMB everything stops; even using SSH it takes a few minutes to return, depending on the size. (Comparison, with all dockers terminated and only a parity check occurring the temperature is 57C/135F. Performance seems substantially improved as well.) The question is do you believe I'm taking a hit because of the temperature (and automatic CPU throttling), because I'm overtaxing this system (it's no beast, I admit), or because I'm expecting more from my drives than I should? Would I see a substantial benefit from upgrading my motherboard/CPU? I've spent a substantial amount on this project and have a very heated wife objecting to further expenditures, but said wife is also an avid user of the resources made available from the server so it dying would not be good either. Is there any information I'm not looking at that would help inform this decision? Config: Sonarr has main directory pointed to /mnt/user/media/Television. [media is the array, cache is turned off]. Radarr has main directory pointed to /mnt/user/media/Movies. [same as Sonarr] Deluge has the download directory pointed to /mnt/user/downloads/deluge [downloads is the share, share is set to only use cache]. NzbGet has the download directory pointed to /mnt/user/downloads/nzbget [same as Deluge]. Sonarr and Radarr move from the downloads directory to the media directory upon completion. All of the rest of the docker configurations are stock, I'm using the linuxservers dockers. One thing I've noticed is that my drives seem to be mostly connecting at 3.0Gbps (SATA II) rather than 6.0Gbps (SATA III). The drives which are attaching at 6.0Gbps are connected to the system with an PEXESAT3221 PCIe SATA III (ASM 1061) controller, so the motherboard ports are not running at 6.0Gbps - so that's one upgrade which would occur. But I've been led to believe that SATA III vs II is really irrelevant because a drive can't saturate the bandwidth anyway. (Parity runs at 165Mb/sec, which I think is pretty good.) I normally have -- max -- three streams feeding Plex. Normally two. Another benefit from a motherboard upgrade would be additional PCIe slots in which I could add additional SATA controllers. I have physical room for two more drives in the case but no open connections. Even if I did upgrade, it wouldn't be a best of breed situation. At best, I'd be looking at a couple hundred to throw at the situation. Probably an AMD Ryzen 5 2600 and an accompanying ASUS motherboard. (https://cpu.userbenchmark.com/Compare/AMD-Ryzen-5-2600-vs-AMD-FX-6300/3955vs1555) Thanks for your time.
  2. In case anyone doubts that the Mover (or drive access, period) was responsible for this problem I'm now sitting at 148.4Mb/sec parity build speed. Which is from the GUI, but... heh.
  3. Talonius

    Failed Drives

    Brand: Seagate Model: Barracuda ST8000DM004 (from STEB8000100) Age: ~3 months (purchased 12/21/2018) Warranty: No, but only because I shucked it from an external drive case. Short Description: Worked great for a month in my ReadyNAS, then began to have issues writing to disk (for about a month), then failed. No SMART report, but I do plan on using the Seagate tools to see if I can recover it.
  4. Magic? Little fairy dust, frog tongue, mushroom caps, throw it at a pixie and wallah, you've got a magic controller which can take disks offline from unRAID long enough to benchmark 'em. (OK, OK, I admit I was figuring stuff out as I posted the message. It's like talking to a duck. Until you explain it to the duck, it's just not clear.) The Mover's running! AHHHHHHHHHHHHHHHHHHH So... excuse me while I go sit in the corner. Thank you gentlemen. I apologize for wasting your time.
  5. Hi all, I have a very low parity build speed. Might be because my disks are trash, I'm not an expert, I tend to buy what I can. Or possibly the motherboard is trash. All dockers and applications have been stopped except for the DiskSpeed plugin. (And to those of you who make these incredible plugins, You. Are. Gods.) Details of the parity build: Total size: 10 TB Elapsed time: 23 hours, 40 minutes Current position: 3.79 TB (37.9 %) Estimated speed: 25.1 MB/sec Estimated finish: 2 days, 20 hours, 39 minutes Motherboard: ASUSTeK Computer INC. - M5A78L-M/USB3 (https://www.asus.com/us/Motherboards/M5A78LMUSB3/specifications/) CPU: AMD FX™-6300 Six-Core @ 3500 Memory: 32 GB Network: eth0: 1000 Mb/s, full duplex, mtu 1500 Slots: PCIe x1, PCIe x16 SATA Controller: ASM1062 Serial ATA Controller (http://www.asmedia.com.tw/eng/e_show_products.php?cate_index=117&item=151) SATA Controller: [AMD/ATI] SB7x0/SB8x0/SB9x0 SATA Controller [AHCI mode] sdh (disk 2) and sdi (disk 3) are Seagate Ironwolf Barracuda drives, 7200 RPM, 10TB. But their benchmark speed in the Diskspeed Plugin is crap. Is that because the parity rebuild is occurring at the same time? (I'm hoping.) These are both attached to the ASM1062 SATA card (and signaling 6Gbps.) The rest are connected to the onboard SATA (which is in AHCI mode, at least): I have set the tunables for larger arrays as instructed in other posts in the forum: The rest of the drive details: The two blank 10TB drives are shucked from WD External Drives (https://smile.amazon.com/gp/product/B07G3QMPB5/ref=ppx_yo_dt_b_asin_title_o06_s00?ie=UTF8&psc=1; they were $149.99 each.) Help? Please let me know if there's any additional information I can give you to help figuring this out. Thanks. Side note: I updated the BIOS on the motherboard to the 2101 revision and that resolved an issue I was having where the box would completely shut down on me, in case anyone else has an M5A78L and running into that problem.