brainbone

Members
  • Posts

    272
  • Joined

  • Last visited

  • Days Won

    1

Everything posted by brainbone

  1. Exactly. Hence my concern. I've resorted to taping up the vents on the case and keeping all drives spun up. Currently have them at around 55°F (~ 13°C). I'll need to find a permanent solution, though my preferred solution would be to shut off select (in my case, all) chassis fans based on HDD temp, however keeping the array spun up is also needed. An unraid feature to force spinup when HDD temps are too low would be nice, however since temperature isn't available when they are spun down, it could be difficult to implement/use, requiring at least one HDD or an SSD be kept spinning as a canary. Low temp alerts in addition to high temp alerts would also be nice. (Guess I'll go create a feature request for these.)
  2. Just like we used to do on extremely code days for car radiators eh? (I'm now using a sheet of plastic, since cardboard is too flammable to have that close to a boiler -- but the temps are still are still pretty cold @~40°F. I'll probably need to add some tape!) I'd like to avoid adding more hardware, but you're probably right. I may need to add a separate fan speed controller.
  3. Any way to get Autofan to control more than one PWM controller and/or fan? I run one of my servers in a boiler room, next to the fresh-air intake. (Yeah, Yeah, but it is what it is.) This works great most of the time, since when the boiler runs, the make-up air from the fresh-air intake nicely cools down the server, offsetting the heat from the boiler room. Problem is, when it's -27°F (~ -33°C) outside, like it is today, the HDDs are running at 19°F (just spun up) to 28°F (spinning) (~ -7°C to -2°C), which is probably way too cold to run these HDDs. I have multiple chassis fans. I'd like to automatically spin all of them down when it's this cold. If any chassis fan is running, even the exhaust fans, too much cold air is pulled in.
  4. The is a critical enough issue where you really should remove 6.6.4 from the stable release category.
  5. If you're not running a cache drive mirrored, yet expect data on the cache drive to be flushed to your protected drive pool nightly, but then a week down the line your unprotected cache drive dies, everything you expected to have been flushed to the protected drive pool but wasn't is now lost. I would call that data loss.
  6. Not moving data from cache drive to the drive pool as expected could lead to data loss.
  7. Looks like I'm experiencing the same thing. I don't understand the downgrade in priority. I wouldn't call not running mover as scheduled just an "annoyance". That's a pretty critical piece of caching for most. Guess it's back to 6.6.3....
  8. Thanks! Never noticed that before. I'm a bit embarrassed.
  9. Just updated to 6.6.4. The help entries that normally required you to click the item label to expand and see are now all expanded by default. This was not happening on 6.6.3. Can't seem to find a setting to change this behavior. Edit: Looks like if I use an incognito window to log into Unraid, things appear as they should. I suspect if I clear my cookies, cache, etc., it would work as well, but I really don't want to, and shouldn't have to, do that. Example:
  10. Some guests, like pfSense w/CARP (for HA), require using e1000, not virtio, when running in KVM. In Unraid, you must switch to xml view to set e1000. Switching out of xml view will cause the change to e1000 on NICs to be reverted to virtio if saved. Please either add selection of NIC type, or at the very least don't overwrite xml selected NIC type, when in form view.
  11. Running into an issue where the mouse cursor is invisible in noVnc after installing the QXL Graphics Driver on Windows guests. Looks like there is now a "Show Dot when no Cursor" option in noVnc that should solve the issue, but Unraid doesn't seem to have that incorporated in the noVnc it's using: https://github.com/novnc/noVNC/pull/1119
  12. I haven't found a fix, but it looks like Unraid needs to update the noVnc they're using: https://github.com/novnc/noVNC/pull/1119
  13. Looks like it's an issue with Sunday. Monday seems to generate a more correct cron (though I'm not sure what the "|| :" is trying to do): # Generated parity check schedule: 0 0 * * 1 /usr/local/sbin/mdcmd check &> /dev/null || :
  14. I'm having the same issue on 6.6.3. Set to run weekly, but it's running daily. My 6.5.3 is still running weekly fine. I've tried multiple changes to the scheduler with no luck. # Generated parity check schedule: 0 0 * * * /usr/local/sbin/mdcmd check &> /dev/null || :
  15. Not sure what's causing it, but it's something to do with the newer versions of LE8. (maybe the newer nvidia driver?) I was able to get it working with the one from 2017 in the archive folder. However, since my main goal with this VM is retro game emulation, I've given up and installed Ubuntu (had to use 16.04, 18.04 didn't work) and RetroPie, allowing me to more easily keep up-to-date.
  16. Agreed. Please submit a feature request too!
  17. Yes, and I did submit a feature request -- but a post here lets other know my reasoning for wanting it back, should they also see fit to request it.
  18. Thanks for the guide. Thought I got it all installed correctly, but when I select a rom, it acts like it's going to load it, but then just returns to the rom selection screen. Here's my es_log.txt: Videos seem to play fine, but can't get a game to load. Any thoughts?
  19. But, unfortunately, doesn't tell me what it would draw at idle. I probably wouldn't be passing anything through to a VM, though I was looking over the threads on how to do it - mainly in hopes that the guest OS may be able to better put the GPU in a lower power mode than unraid seems able. That''s exactly my concern. For whatever card I choose, will unraid be able to keep it in a low power state.
  20. So, after deciding to go with a Ryzen 2700X system (a choice I'm somewhat regretting, kicking myself for not going with server-grade hardware, but here we are), I'm trying to find the lowest power discrete GPU. All it'll be used for is local console output. I've been looking at the ZOTAC GeForce GT 710 PCIe x1, so I don't need to sacrifice a PCIe x16 slot. Anyone have experience with this card in unraid? How many watts does it pull when unraid seems to lack the ability to use DPMS? Any way to put GPUs into a low power mode in unraid?
  21. I'm a bit surprised by Unraid not seeming to support dpms. "Blanking" the screen with dpms can considerably reduce the power consumption for some graphics cards. It should, at the very least, be an option. Just turning off the monitor doesn't have the same effect.
  22. Thanks! The Marvel 9230 controllers look interesting, and I almost purchased two of them, but I'm worried about the virtualization issues... Guess I'll post in that thread to see if there's been any resolution.
  23. Anyone have recent information on power consumption of 4 and/or 8 port SATA / SAS controllers? I'm trying to find something with very low idle power consumption, but having difficulty finding accurate, up-to-date data for various controllers. According to this (old) chart, some SAS2008 based controllers consume ~6.5w idle. Is this as good as it gets for Unraid compatible controllers, or are there better options out there?