Squazz

Members
 View Profile  See their activity

  • Posts

    137

  • Joined

  • Last visited

 Content Type 

Posts posted by Squazz

    Ryzen Build - Gamer & NAS / server

    in Unraid Compulsive Design

    Posted

    On 5/9/2019 at 11:09 PM, tjustice86 said:

    How are you passing through your audio? I would like to use HDMI passthrough but can't seem to get it to work with my VM. The on board audio controller works but I would like to watch movies and even game at at least 5.1. Any ideas or pointers?

     

    MOBO: MSI B450 Gaming Pro Carbon AC

    CPU: AMD Ryzen 7 1700x

    GPU: XFX AMD Radeon rx 580 

     

    I will have to check my cstates and memory speed when I get home.

    Hi there, I'm SO SORRY for replying so late.

     

    Right now, I'm using a USB DAC for my audio. Works like a charm :)

    When at my computer, I'm only running a pair of headphones, no surround.

    [Support] Djoss - dupeGuru

    in Docker Containers

    Posted

    7 hours ago, Squid said:

    And there will never be.  Problem is that when reading the file from /mnt/user/share/..., unRaid will give you the version from the lowest numbered drive (cache, disk1, disk2, etc)  The version which you may be expecting, or which may be the latter version could be on the highest numbered drive.  Hard to say, and I don't want the plugin to attempt to guess what the user wants to actually do.

     

    You can manually clean them up via Krusader (if you map /mnt instead of /mnt/user in the template) by navigating to the appropriate drive(s) and inspect the files accordingly

     

    Should be noted that under normal circumstances, duplicated files do not happen under unRaid.

    I totally understand, and I only have this problem due to me using unBalance :/

    I was hoping there was a way to tell a program "if you find any dupes on diskX, delete them", so that I was in charge of what files was removed. Guess I'll have to do it the hard way :P

    [Support] Djoss - dupeGuru

    in Docker Containers

    Posted

    On 5/5/2017 at 8:45 PM, jonathanm said:

    Very definitely possible, and can be quite problematic for someone who doesn't have a good handle on what a user share is.

    Consider the following.

    /mnt/disk1/myshare/test.txt

    /mnt/disk2/myshare/test.txt

    /mnt/user/myshare/test.txt

    In this example, the file accessed at /mnt/user/myshare/test.txt will be the one on disk1, for all intents and purposes /mnt/disk2/myshare/test.txt doesn't exist in the user share system. However... if you delete /mnt/user/myshare/test.txt, then the one on disk2 will show up the next time the user share system is reloaded. It's quite possible to have 2 or more identically named files this way, but the contents not be equal. You could theoretically hide files from the user shares this way, and cause all sorts of confusion.

     

    Typically this kind of shenanigans occurs when someone with a poor grasp of how unraid works starts playing around with disk shares. That is one of the main reasons disk shares are disabled by default.

     

    In my case, this happened when using unBalance to scatter files from one disk to multiple others.

    As I see this docker could not help with looking for duplicates on the disk-level, so we know of any other docker og app that can help us with finding and deleting these duplicate files? :)

    Docker service wont start

    in Docker Engine

    Posted

    1 hour ago, bonienl said:

    In Unraid 6.7 we do not suppress PHP warnings anymore.

     

    Please follow the advice of @johnnie.black and delete/recreate your docker image. You will need to re-install your containers, if you have CA installed, you can use its restore function to quickly get back all your containers.

    I'm not sure how to use the restore function. Where should I look?

    Docker service wont start

    in Docker Engine

    Posted

    3 minutes ago, johnnie.black said: 
    
Feb 10 16:21:58 NAS kernel: BTRFS warning (device loop2): csum failed root 5 ino 15295 off 0 csum 0x98f94189 expected csum 0xd349d779 mirror 1

     

    Docker image is corrupt, delete and recreate.

    How would I go around doing this? :)

    Docker service wont start

    in Docker Engine

    Posted

    6 minutes ago, CHBMB said:

    Well, you're using a release candidate, so probably should report it there.

    What version did you update from?

    I upgraded to the release candidate because I had the problem before on version 6.6.6
    Upgrading to the release candidate however gave me the error message, that I didn't have before

    Docker service wont start

    in Docker Engine

    Posted

    5 hours ago, CHBMB said:

    Not without more context.  Need to see what you're doing to generate that error message.   To clarify, is it the docker service that won't start?  Or a docker container?  If it's the former, you need to post diagnostics, if it's the latter then some more info on the container etc

    My bad :) It went a little fast with that post ;)

    It's the service that won't start

     

    Diagnostics: nas-diagnostics-20190210-1626.zip

    Image where I got the error-message

    image.thumb.png.306a3c51b8601e4e4dffd13c8663275b.png

    Docker service wont start

    in Docker Engine

    Posted · Edited by Squazz

    I get the following error message

    Quote

    Warning: stream_socket_client(): unable to connect to unix:///var/run/docker.sock (Connection refused) in /usr/local/emhttp/plugins/dynamix.docker.manager/include/DockerClient.php on line 654
    Couldn't create socket: [111] Connection refused
    Warning: Invalid argument supplied for foreach() in /usr/local/emhttp/plugins/dynamix.docker.manager/include/DockerClient.php on line 826

    Warning: stream_socket_client(): unable to connect to unix:///var/run/docker.sock (Connection refused) in /usr/local/emhttp/plugins/dynamix.docker.manager/include/DockerClient.php on line 654
    Couldn't create socket: [111] Connection refused
    Warning: Invalid argument supplied for foreach() in /usr/local/emhttp/plugins/dynamix.docker.manager/include/DockerClient.php on line 890

    Does anyone know what that means and what I should do about it?

     

    Windows issues with unRAID

    in General Support

    Posted

    On 12/5/2017 at 2:36 PM, Frank1940 said:

    Apparently there is another round of problems with Windows 10 computers (which have updated recently) having difficulties connecting to an unRAID server using SMB.  You might want to look here fro a possible solution.

     

            https://support.microsoft.com/en-us/help/4046019/guest-access-smb2-disabled-by-default-in-windows-10-server-2016

     

    There is a possibility that the Samba group may have a fix for this in the near/distance(?)  future.

    In 2019 this saved my day

    Ryzen Build - Gamer & NAS / server

    in Unraid Compulsive Design

    Posted

    8 minutes ago, ksignorini said:

    How do you find your gaming FPS with Ryzen on unRAID?

    This is something i will have to dig deeper into at some time.

    Generally I havn't noticed anything alarming. But for some reason it seems CS:GO is capped at 150FPS when I'm playing it on my Win10 VM on Unraid. If I boot up on barebones I get 280+ FPS

    But this is the only game I have experienced it in (and the only game I play I generate 150+ FPS in)

    Sorry I can't give a better answer at the time of writing this

    Ryzen Build - Gamer & NAS / server

    in Unraid Compulsive Design

    Posted

    I'm glad that my updates to the post is helping someone, it makes it worth the work! :)

    After your kind words, I'll continue updating the post with new info

     

    14 hours ago, tsawind said:

    So I just re-read part of this post, and I think you just explained why I have been having BSODs. They have been: attempt_to_write_to_read_only, memory_management, dxkrnl.sys, netkvm.sys, ect ect ect.


    I have an overclocked build with a 2700x.  I bought 2 8gb sticks of good high end RAM, G.SKILL Trizent Z 3600 CAS 1.  I have them running at 3600 mhz, command rate 1, Timing 17-18-18-38, 1.35v.  I have noticed during overclocking the speed and timings that these particular sticks don't want to run faster than 3800mhz, and they like extra voltage to improve timings.  From an overclocker standpoint they only went unstable with too high of speeds, or if the temps went over 50 degrees C.

    ...

     

    I am going to reduce the speed, and check the max voltage allowed to the memory controller for my ryzen cpu.
    https://en.wikichip.org/wiki/amd/ryzen_7/2700x

     

    From what I see here I have "G.SKILL TridentZ Series 16GB (2 x 8GB) 288-Pin DDR4 SDRAM DDR4 3600 (PC4 28800) Intel Z170 / Z270 / Z370 / X299 Desktop Memory Model F4-3600C17D-16GTZ", which appears to be single rank.  I should run this ram on this particular build at 1.35+- volts, and DDR4-2133.  I believe I should wait to do the major timing "overclock" until I have all 4 DIMMS installed?

    It takes forever, and the only way I know how to do it is, bare-metal boot, turn down a timing number, stress test, rinse and repeat.
    Anybody see anything I am missing??

    After my experience with BSODs and parity errors I'm not overclocking my memory anymore. The gains vs. the risks are not worth it to me.

    But yes, if you want to do a overclock, you should wait until you have all the sticks. Everything would have to be re-tested and probably reconfigured when you get the last sticks.

     

    14 hours ago, tsawind said:

    I bought a SanDisk Cruzer Fit CZ33 32GB USB 2.0 Low-Profile Flash Drive- SDCZ33-032G-B35   I am hoping this is a decent stick for my UNRAID.

    My build: https://pcpartpicker.com/list/tJ39QZ (some misc parts are missing like custom case, a USB controller, USB hubs, extension cables ect.)

     

    ...

     

    Looking for answers to:

    1. Is that a proper USB stick?

    From what I have understood, all USB sticks are fine. I'm running mine on a stick I got for free at a conference, works just fine even though it's a "slow stick".

    The same sector keeps being corrected in subsequent parity checks

    in General Support

    Posted · Edited by Squazz

    On 12/3/2018 at 12:45 PM, johnnie.black said:

    It's a possibility and it doesn't hurt to try, there were at least of couple of similar cases with other Ryzen users where parity sync errors stopped after RAM clock was lowered.

     

    On 12/3/2018 at 3:48 PM, John_M said:

    Your sig says "2x 8GB 3200Mhz". Is that just the spec of the RAM or also the speed at which you're running it? If the latter then it's way out of spec for a 1st gen. Ryzen. What are the part numbers for the DIMMs? The fastest you can run the memory controller is 2666MHz without overclocking it but depending on how may DRAM chips you're hanging on the bus you might have to de-rate it further.

     

    I lowered the speed to 2400MHz, and have now made two parity checks. The first run found a single error. The second run didn't find anything.
    I have yet to confirm that the errors stops coming, but for now it seems that it helped.
    Thanks for everything so far :)

     

    Edit: Yes, the errors stopped after lowering the speed

    The same sector keeps being corrected in subsequent parity checks

    in General Support

    Posted

    9 minutes ago, John_M said:

    There are ten entries for that part number in the table you referenced. Some of them show dual rank/double sided and some show single rank/single sided, depending on whether they use 16 x 4 Gb or 8 x 8 Gb DRAM chips. For example:

    and

    It isn't simply a difference between the Ripjaws V and Trident Z brands - both come in both configurations. The part number you gave me is too generic. When I checked on G.Skill's site it said single/single too. Presumably it only shows the newer type with the bigger chips.

    Oh, my bad.

    It's the F4-3200C16D-16GVKB version I have

    The same sector keeps being corrected in subsequent parity checks

    in General Support

    Posted

    8 hours ago, John_M said:

    I haven't been able to confirm for your specific motherboard* but two DIMMs of that RAM should run at DDR-2666 without taking the memory controller of an R7 1700 out of spec. With a 2000-series CPU or APU you would be able to run it at DDR-2933.

     

    *But I have been able to confirm it's the case with all the others I was able to check, including the ASRock X370 Pro4 that you mentioned.

    For Dual Rank memory, the memory controller on the Ryzen 1700 is only supporting 2400Mhz

    https://en.wikichip.org/wiki/amd/ryzen_7/1700#Memory_controller

    The kit I have, is dual rank memory: https://www.reddit.com/r/Amd/comments/649ay8/ram_collection_thread_please_post_your_ram/

    The same sector keeps being corrected in subsequent parity checks

    in General Support

    Posted

    43 minutes ago, John_M said:

    Your sig says "2x 8GB 3200Mhz". Is that just the spec of the RAM or also the speed at which you're running it? If the latter then it's way out of spec for a 1st gen. Ryzen. What are the part numbers for the DIMMs? The fastest you can run the memory controller is 2666MHz without overclocking it but depending on how may DRAM chips you're hanging on the bus you might have to de-rate it further.

    They are running 3200Mhz, running that speed due to XMP.

    The part number for the memory is F4-3200C16D

     

    43 minutes ago, John_M said:

    I'm interested to hear of any AM4 motherboard that actually supports ECC as I haven't found one yet. Yes, many allow you to use ECC RAM and some even have the necessary extra tracks to the sockets, but I'm not aware of any that have support for ECC in the BIOS. It's a shame because the processors all support it but the motherboards are aimed at gamers. Perhaps if SuperMicro or ASRock Rack ever develop an AM4 board... They both make EPYC motherboards though, so I think I'm going to use one of those in my next server.

    Asrock is making their PRO series that supports and runs ECC memory: https://www.asrock.com/mb/AMD/X370 Pro4/index.asp#Specification

    The same sector keeps being corrected in subsequent parity checks

    in General Support

    Posted

    1 hour ago, johnnie.black said:

    Memtest can't prove a negative, i.e., you can't be 100% sure memory is fine when no errors are detected, it can only prove there are problems when an error is detected, hence why I use and recommend using ECC RAM for files servers if you care about data integrity.

    I'd love to run ECC, but I was stupid enough to buy an MSI board for my Ryzen 1700. It can boot ECC memory, but ECC is disabled.

    Regarding memTest and the errors.
    If I can't find errors on the memory on any frequency, then it could still be a frequency problem?

    I'll have to try and give parity checking a few more runs. And double check if I consistently can get rid of the errors with a lower frequency