Jump to content

Kaldek

Members
  • Content Count

    13
  • Joined

  • Last visited

Community Reputation

1 Neutral

About Kaldek

  • Rank
    Member

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. I suppose it depends how busy your server is but I personally don't pin Docker containers to CPUs. My CPU is an i7-6950X though so I also don't know the per-core processing capabilities and whether the audio transcoding impacts this CPU much. TL;DR I haven't noticed any CPU consumption even with hardware decoding using a GPU. Also, most of my streaming is to devices that support direct audio and don't transcode, so I should probably monitor next time I'm streaming to a device like an Android phone etc. Having said all that, the cores on my server are mostly idle unless I'm doing actual work on the server using VMs. Even 50% CPU on a few cores wouldn't get my attention.
  2. Hi everyone, I have recently had the same issue of disks not spinning down. I seem to have managed to solve this (I believe) but I wanted to document what I believe the causes were. Whenever disks were not spinning down I could see a consistent ~3.4KB/s read from the disks in question. I could not see what was reading these disks as it was not listed in "Open Files" or "File Activity" tools/plugins that I have installed. So in my case what it looks like the cause was that I had shares which I had reconfigured to exist only on one disk, or to live purely on the cache drive (cache "only"), but the file allocation system of unRAID appeared to not be fully respecting this. What it looks like is that even if there is a blank directory on a disk that previously contained components of a share, then any disk writes to that share still keep probing that disk and keeping the disk spun up. The solution in my case was to do the following: Shut down all VMs and Docker containers Edit all shares that need to only be on the cache to "Prefer" for the cache setting Run the Mover Edit the same shares again to use "Only" for the cache setting Review all shares that should only be one some (or one) of the disks and make a list of these Download and install unBALANCE and run it Use unBALANCE to "Gather" any of those shares that should only be on specific disks, and place those shares only on the disk targets that I wanted them on Stop unBALANCE Spin down all the disks I know should be spun down during idle times Start all the VMs and Docker Containers I need running After that, I confirmed that my disks stayed spun down. For the past 30 minutes, I've not seen any of these disks spin up so it looks stable (so far). Note that all my VMs and containers run off the SSD cache which is a 1TB mirror. You might ask "why care about this"? Well since my server is running 24/7 the difference in power consumption between the disks being spun up and not being spun up is about 35 watts. This seems "like nothing" but power is not cheap here in Australia. During peak times we pay 38c cents per kw/h. Every little bit of my base load ends up going towards a yearly power bill of $1,500 and that's after all the rebates from my 5kw/h solar array (which pays me 12c kw/h for exported power). My entire server rack - at idle now - pulls 151 watts. Two weeks ago that idle was 240 watts, which is a near 90 watt saving. I've gained about 40-45 watts of this saving from unRAID tuning. The rest came from replacing outdated Unifi UAP-AC access points with newer UAP-AC-LR access points that pull much less power.
  3. Finally I'm getting to make use of all these cores on my 6950X.
  4. I know there are other posts on this topic, but none of them appear to solve or document the solution. My situation is this: I have a dual stack IPv4/IPv6 Internet connection My ISP issues me a /56 IPv6 prefix I use a Mikrotik router, which notifies hosts in the network about IPv6 via SLAAC, *not* DHCPv6 Every host in the network, including unRAID on br0 is getting an IPv6 address The result of "docker inspect br0" has IPv6 set to disabled (and I can't see how to change that in the GUI) Here's the result of ifconfig: Here's a result of a ping: Here's my network settings: Here's the docker settings: Here's the result of "docker inspect br0":
  5. Upgraded successfully from 6.8.1 to 6.8.2. I waited for the unraid Nvidia build for 6.8.2 to be released. Everything working fine.
  6. Upgraded to 6.8.2. Everything working including HW transcoding nicely on my Quadro P400.
  7. No apps or scripts writing to the flash drive. If it's hardware then it will have to be the motherboard. I've got other devices connected via USB though that keep working (Logitech Unifying receiver, and external USB 3.0 8TB disk). Only the flash drives die. They've been in multiple ports, but all USB 3.0 ports so far. For reference, the motherboard is an EVGA X99 Micro2 with an Intel i7-6950X.
  8. No, I'm blaming the OS for locking its license to hardware IDs. Australia is the ass end of the world when it comes to support from overseas and when you need your server all the time, having to wait for a license reset or buy a new license isn't suitable when your vendor is asleep for the next 8 hours, and then working through the list of other support tickets for the following four hours. My server potentially having a USB hardware port fault wouldn't cause a problem in relation to unRAID if I was able to rapidly recover to a new USB drive and use on an online license authorisation.
  9. My unRAID 6.7.2 server has so far killed three Flash Drives. Each time, the server will keep running fine with no alerts or warnings but as soon as I reboot I will get either: "Kernl panic - not syncing: VFS: Unable to mount root fs on unknown block(0,0)"; OR "RIP: 0010:panic+0x1e6/0x227" On one instance, I managed to reformat the flash drive and copy back the flash backup I had, and it recovered. On the other two occasions (including the most recent), only replacing the USB flash drive and restoring my config to that new drive fixes it. Unfortunately, because I need my server 24/7, the last time it died I was stuck with no license. As this was the second key regeneration, I had to raise a support ticket - which I cannot wait for - and instead had to buy another license because I needed my server now. This is getting beyond a joke. Where is the online license option where there can only be one instance of a server running? I can't wait for support to fix issues when I'm on the other side of the planet. I need my licenses to work - always - regardless of whether I have just had to recover from yet another flash drive failure. At least allow more than one key reset per year without having to raise a support ticket!
  10. As above, in the flash settings. This is all related to SpaceInvadorOne's video on passing a NIC through to a VM:
  11. Ugh, I'm such an idiot. It's because I booted to GUI mode and I didn't have the "vfio-pci.ids=10ec:8168" after the append command. *Sigh*
  12. I have an unraid 6.7 system based off an X99 motherboard and i7-6950X with 64GB of memory. This system has an onboard r8169 ethernet NIC and a PCI-e TP-Link branded e1000 NIC. The PCI-e card is in its own IOMMU group and is used as a sniffer NIC attached to a VM (which runs NtopNG) and the cable is connected to a mirror port off my network switch. The onboard NIC has been configured to be eth0 for...well...forever, and the PCI-e card is eth1. Today, I rebooted and eth0 is now eth1, and of course eth1 is now eth0. I cannot for the life of me work out what the hell is going on. Dmesg of my system is attached. Note this dmesg includes when I switched the cables around and allowed the PCI-e card to operate as eth0 so I can at least use my blimmin' server. I need to FORCE the onboard NIC to *always* be eth0 no matter what. I cannot rely on this renumbering. What happens when it decides to renumber itself again? dmesg.txt
  13. Hi folks, my flash drive stopped allowing disk writes (I got no notifications of this) but none of my docker images appeared to be updating. So I rebooted my server and have the dreaded "Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block (0,0). The flash drive is borked. I can't even repair it on Windows. I do not have a flash drive backup from anywhere recent, but I DO have the config folder from the flash drive (minus two plugin TGZ files which refused to copy). I need help on how to install a new flash drive but import the config folder I have from the old flash drive. I really can't have this cause me to lose all my data and my VMS. Please help.