Jump to content

song

Members
 View Profile  See their activity

  • Posts

    129

  • Joined

  • Last visited

 Content Type 

Posts posted by song

    万兆网卡使用iperf3测速可以满速，但是拷贝文件经过缓存设备时只有5g速率

    in Chinese / 简体中文

    Posted

    22 hours ago, wangxun said:

    我unraid和电脑都是跟你一样的万兆卡  直连或者万兆交换机都不行  各种办法都试过了 不行 我现在的情况是电脑传一个30G的电影到NAS  只有200M/S  但是NAS传到电脑有800M/S

    检查下你nas插万兆网卡的pcie速度是多少，这个卡是pcie2.0 x8我目前unraid是主板第一个插槽pcie 3.0 x16速率是满速的，unraid使用一个口，另一个口直通给群晖，群晖可以满速，还要检查你的电脑pcie速度是多少，实测pcie 3.0 x4速率可以跑满这个卡的万兆速率，检查下你的pcie是不是x2有些主板是可以在bios内改成x4的但是需要禁用一个或者2个sata接口，如果是x2是跑不满的，另外如果是x4跑不满还要检查你的光模块，我之前使用两条dac线把群晖和unraid连接万兆交换机，电脑使用aoc线也是跑不满，可能是不兼容，后来全部更改为aoc光纤线后都是满速，iperf3在unraid测速是满速的只是拷贝文件不是满速，群晖没这个问题

    万兆网卡使用iperf3测速可以满速，但是拷贝文件经过缓存设备时只有5g速率

    in Chinese / 简体中文

    Posted

    10 hours ago, JackieWu said:

     

    有没有尝试过避开 SHFS 文件系统进行读写测试？

     

    参考：为什么unRAID下载/拷贝文件时CPU占用高：详解FUSE-SHFS、I/O问题和解决办法

     

    Snipaste_2024-01-26_01-02-11.thumb.png.157e8669be77f4d2d1cd51ce7452bccd.png

    没有这样尝试过，但是在复制文件的时候，cpu占用确实会上去，iperf3测速的时候cpu几乎没占用，如果使用这个办法，他还会把缓存内的文件迁移到我想要移动的文件夹内吗？因为迁移文件后，缓存内实际上是没有目录的，只有在执行拷贝文件时，缓存才会创建相同的文件路径，如果使用这种办法是不是需要我手动移动文件到指定的文件夹内？

    万兆网卡使用iperf3测速可以满速，但是拷贝文件经过缓存设备时只有5g速率

    in Chinese / 简体中文

    Posted · Edited by song

    23 minutes ago, JackieWu said:

     

    如果你那里解决了这个问题，希望能在分享出来，我自己没有万兆网卡，所以希望能多了解一些情况。而且现在中文板块的活跃度还比较低，多解决一些问题也能够帮助到大家。

    测试了 没区别，smb多通道打开关闭都是一样的速率

    Unraid失联到底能不能解决啊？好多人反应了

    in Chinese / 简体中文

    Posted

    On 6/4/2023 at 2:53 PM, Lian said:

    本来没入正前不好意思来论坛反应 但我们迟早会入正的啊 白瞎我这么费劲的帮你们安利了 系统哪儿哪儿都好 就是不稳定 这个对nas来说太致命了 你叫我怎么继续吹unraid？不说多精通 确实把unraid的规则弄明白了 确实吊 但这个内核崩溃失联 必须强制物理关机真的太要命了 这些硬盘本来就怕突然断电 你这失联的问题导致我们必须强制关机 是硬盘滞销了想让我们帮帮卖硬盘的吗？能不能解决给句话 中文区的版主 能不能反应上去？尽快修复这个问题 你们要真没能力解决 那也别怪我当黑粉了 失联问题就是你们解决不了 不关用户硬件不兼容啊还需要改什么设置呀的事

    中文板块官方没人理的。我都是翻译去他们的意见反馈版块发，他们官方的人才会理你，有很多的问题，都是他们给我解决的

    Unraid失联到底能不能解决啊？好多人反应了

    in Chinese / 简体中文

    Posted

    On 9/8/2023 at 4:04 PM, xsk said:

    docker如果用的macvlan就有可能出现失联，我的反正改ipvlan就好了。

    从6.9.2开始转正后一直用，从没失联过，只用intel的网卡，螃蟹什么的从来不用，6.12版本的时候升级只遇到过直通的设备id错乱，其实也很好解决，关于失联用了快两年了没遇到过一次，可以说它虚拟机性能对比其他的esxi pve弱。但是失联真没遇到过。一直是用的macvlan没改过

     

    iShot_2024-01-25_23.29.27.png

    万兆网卡使用iperf3测速可以满速，但是拷贝文件经过缓存设备时只有5g速率

    in Chinese / 简体中文

    Posted

    目前使用最新版6.12.6，网卡使用的是华为sp310 82599es芯片双网口万兆网卡，插在主板第一个pcie插槽速度是满速的，使用iperf3作为服务端，苹果下也是同样的网卡，连接的模块为aoc光纤，iperf3测速可以跑满，但是在电脑上拷贝文件到unraid只有5g的速率，unraid使用一块三星970 evo 250g m2固态硬盘作为缓存设备，按理说是能够跑满万兆速率的，不知道为什么只有5g的速度，以下为测速截图

     

    iShot_2024-01-25_23.16.12.png

    iShot_2024-01-25_23.16.43.png

    iShot_2024-01-25_23.17.20.png

    [Support] Fork Unraid API-RE

    in Docker Containers

    Posted

    This is my system log after running the Docker container on 'unraid API'. I tried to open a container, but unfortunately it directly crashed the 'unraid' webui page. I had to force 'unraid API' to be closed to restore it to normal

     

    System log information

     

    Aug 2 11:34:42 song kernel: docker0: port 4(vethb6da9cc) entered blocking state Aug 2 11:34:42 song kernel: docker0: port 4(vethb6da9cc) entered disabled state Aug 2 11:34:42 song kernel: device vethb6da9cc entered promiscuous mode Aug 2 11:34:44 song kernel: eth0: renamed from veth05a0186 Aug 2 11:34:44 song kernel: IPv6: ADDRCONF(NETDEV_CHANGE): vethb6da9cc: link becomes ready Aug 2 11:34:44 song kernel: docker0: port 4(vethb6da9cc) entered blocking state Aug 2 11:34:44 song kernel: docker0: port 4(vethb6da9cc) entered forwarding state Aug 2 11:34:48 song webGUI: Successful login user root from 172.17.0.5 Aug 2 11:34:57 song nginx: 2023/08/02 11:34:57 [error] 6871#6871: *539406 open() "/usr/local/emhttp/plugins/dynamix.vm.manager/templates/images/DS" failed (2: No such file or directory) while sending to client, client: 172.17.0.5, server: , request: "GET /plugins/dynamix.vm.manager/templates/images/DS HTTP/1.1", host: "192.168.50.254" Aug 2 11:35:27 song php-fpm[4970]: [WARNING] [pool www] server reached max_children setting (50), consider raising it Aug 2 11:38:22 song kernel: veth05a0186: renamed from eth0 Aug 2 11:38:22 song kernel: docker0: port 4(vethb6da9cc) entered disabled state Aug 2 11:38:22 song kernel: docker0: port 4(vethb6da9cc) entered disabled state Aug 2 11:38:22 song kernel: device vethb6da9cc left promiscuous mode Aug 2 11:38:22 song kernel: docker0: port 4(vethb6da9cc) entered disabled state

     

    I am unable to obtain the log information of 'unraid API' because opening the container while it is running will cause the webui of 'unraid' to become unresponsive. I must stop 'unraid API' to restore it to normal. However, after stopping the operation, the log information of 'unraid API' will be lost and the log from the last startup will not be displayed. I can only provide you with the system log of 'unraid'

     

     

    [Support] Fork Unraid API-RE

    in Docker Containers

    Posted

    10 hours ago, BoKKeR said:

    @song I have tested the vm start/stop functionality on a test machine, it works on 6.12.3, Can you tell me more about your system, any logs when this happened? Also could you send the HTML files as stated above? 

    My system version is 6.12.3. When this issue occurred earlier, I did not record any relevant information because I have not been using it recently. I remember when this issue occurred, there were a large number of warning messages in the system log

    [Support] Fork Unraid API-RE

    in Docker Containers

    Posted

    Thank you for maintaining this project. The good news is that you were unable to control the virtual machine before. Now, the information about the virtual machine appears in the plugin, but after connecting to the home assistant, the entity still cannot control the Docker container and virtual machine, I tried to control the container and virtual machine on the web ui page of 'unraid apis', but the results were equally uncontrollable, and even made the web ui page of' unraid 'crash and inaccessible, which took a few minutes to access. I don't know what caused it

    [Plugin] NUT v2 - Network UPS Tools

    in Plugin Support

    Posted · Edited by song

    My UPS model is BK650M2, When I upgraded to version 6.11.5, the plugin was running normally and there were no errors in the logs. However, after upgrading to 6.12.3, the old version of the plugin became unusable and I used the new version of "Network UPS Tools". Since then, it has been continuously reporting errors and a large number of warning messages have appeared in the logs. Recently, after upgrading to the latest version 2.8.0 of "Network UPS Tools", the UPS has started to lose connection and must be manually reloaded The 'Network UPS Tools' can only be reconnected after a period of time, but the connection will still be lost. I tried uninstalling and reinstalling' Network UPS Tools', but after a night, it did not disconnect again, but warning messages kept appearing again.

     

    This is a warning message after disconnecting

    iShot_2023-08-01_12_01_43.thumb.png.1d2e8096f971d0e0d9f44985f1382470.png

    This is a constantly occurring warning message

    iShot_2023-08-01_12_03_35.thumb.png.2a7584420db9b9a92c125ff20465b67e.png

    This is the plugin's settings page

    iShot_2023-08-01_12_05_37.thumb.png.c73c8f15378bd6f5eac465923f815adf.png

    Dynamix File Manager

    in Plugin Support

    Posted

    On 6/29/2023 at 10:19 AM, BOGu said:

     

    Yes, I'm having this problem too! I am a Chinese user

    I have also encountered this issue, and I even have to reinstall the system. After reviewing your feedback, I realized that this is a plugin issue. Please resolve it as soon as possible!

    [Support] Electric Brain Unraid API

    in Docker Containers

    Posted · Edited by song

    On 2023/5/10 at PM10点53分, BoKKeR said:

    我已经暂停了这个项目，因为我相信 unraid 团队应该很快发布他们的 graphQL api。发布的 quickfix 仍然有效

     

    I would like to know where you got the news. The "unraid" team seems not interested in their API connecting to home assistant. They are currently focusing on developing their "unraid net", which is not meaningful for me personally. I would rather "unraid" appear on my home assistant dashboard, I am actually looking forward to one day not having to use third-party tools to access the home assistant to monitor my "unraid" host status and control virtual machines and containers, but I think this is very far away!

    Can a system that can't work properly with the most basic functions be called a system?

    in Forum Feedback

    Posted · Edited by song

    26 minutes ago, ich777 said:

    Then may I ask why you link to the post?

     

    This was definitely a issue with my plugin and was only a issue when the LXC plugin was installed but after I was informed about that I looked into that and updated the plugin so that this won't happen anymore in a matter of days IIRC. But please keep in mind I'm a community developer and not related to Unraid.

    Didn't you see the keyword "Shared folder disappeared"? I declare again! I did not install "lxc plugin"!!!

×
×
  • Create New...