Ryzen here. Updated from rc16 to rc18. All containers and VMs (CentOS 7, Ubuntu 16.04, Windows 10 + GPU passthrough) started automatically. Restarting a VM works fine. Once again, the system detected unclean shutdown and triggered parity check.

Edited December 30, 20178 yr by realies

Looking good so far - rebooted two W10 VMs successfully

 

Re the new terminal - is the session persistent between browsers/machines, or do I need to use screen like I do with Putty?

6 hours ago, realies said:

Ryzen here. Updated from rc16 to rc18. All containers and VMs (CentOS 7, Ubuntu 16.04, Windows 10 + GPU passthrough) started automatically. Restarting a VM works fine. Once again, the system detected unclean shutdown and triggered parity check.

I’ve found if I manually turn off my VMs and reboot for a software upgrade that issue goes away. (Also using Ryzen)

Quote

• We now have a web-based Terminal application available by clicking the Terminal button on the Menu bar.  Kudos to @eschultz for integrating the ttyd application with our nginx install.  This is using websockets and appears to be quite fast, give it a try!

Dang I like this feature!  I can now work in a bash session without having to be on the one computer that is a VM that I have Putty installed.  It's nice because I can't always get to the VM.  For example if the array is stopped.

12 hours ago, Ashe said:

the terminal does not show up in black and white themes only azure and grey themes, has anyone else seen this?

 

Terminal works with the black theme here (Chrome/Mac)

24 minutes ago, JustinAiken said:

 

Terminal works with the black theme here (Chrome/Mac)

Also for me.

One thing I noticed about the new terminal..

 

# From ssh or telnet:
$ echo $HOME
/root

# From the new web terminal
$ echo $HOME
/

 

..which means that `/root/.bash_profile` doesn't get loaded in the webterminal, so you have to `source /root/.bash_profile` if you want the nice pretty prompt

Edited December 30, 20178 yr by JustinAiken

2 minutes ago, JustinAiken said:

One thing I noticed about the new terminal..

 

# From ssh or telnet:
$ echo $HOME
/root

# From the new web terminal
$ echo $HOME
/

 

..which means that `/root/.bash_profile` doesn't get loaded in the webterminal, so you have to `source /root/.bash_profile` if you want the nice pretty prompt

 

Yeah we fixed that too in next release.

On 29.12.2017 at 12:48 PM, bcjenkins said:

He's referring to this: 

I had this same problem in 17b and did not have it when I upgraded to 18f. 

 

hih

 

 

Yes, indeed.

 

Can confirm, the rc17b issues are gone (upgraded from rc15).

 

My Docker IPs have stayed intact!

installed today on a z400 with the purpose of doing 2 computers 1 pc. Works without issue.

On 12/28/2017 at 12:39 PM, jonp said:

 

Correct.  It's untenable for us to support AMD specific issues with a special kernel build.  The purpose of the rc17b separate bzimage file was to confirm whether or not the patch even worked for those users to see if they were at least on to something.  Unfortunately the "ugly patch" is ugly simply because of the fact that it can break stability for other users that aren't on AMD platforms.

 

The unfortunate state of affairs with AMD right now is that they have some Linux-specific issues they need to address before Ryzen / TR can be truly recommended.  I love the idea of a high-core count low cost CPU solution that competes with Intel, but it has to be feature-complete and stable first.

Still going to support FX-xxxx CPUs and AM3+ Motherboards? in the 6.4 releases?

2 minutes ago, Paul_Ber said:

Still going to support FX-xxxx CPUs and AM3+ Motherboards? in the 6.4 releases?

 

Yes, older AMD still seems to work.  It's just the new zen platform that's going through some growing pains right now.

Just now, eschultz said:

 

Yes, older AMD still seems to work.  It's just the new zen platform that's going through some growing pains right now.

With the upgrades, lost the good video card to family member.  Built another PC(unRaid box next to desk) with a Quad LCD mount, so VMs not really needed right now.  Don't try to overclock a video card in a VM, you will crash your whole system.

I have my default shell set to zsh but the web terminal appears to default to /bin/sh. Why won't it default to the user's set default shell?

 

Also, I hadn't posted about this previously but the CPU meters on the Dashboard no longer work on Mobile Safari. Does this have to do with the use of Web Sockets?

23 minutes ago, Taddeusz said:

I have my default shell set to zsh but the web terminal appears to default to /bin/sh. Why won't it default to the user's set default shell?

 

Next release will prob solve that

 

24 minutes ago, Taddeusz said:

Also, I hadn't posted about this previously but the CPU meters on the Dashboard no longer work on Mobile Safari. Does this have to do with the use of Web Sockets?

 

Yes there's an issue with Safari and websockets.

if u can fix the auto restart error for windows 10 on a laptop while using the unraid os server on the next version, i tried 4 cpu cores, then 3, and 2, still the same error = windows 10 auto restart 

I would like to request that the Tehuti drivers that were added in rc16 be removed as it seems to be causing boot issues and they don't seem to be the proper drivers for any of the new Tehuti networking chips. Hopefully removing legacy support due to it causing issues for users of the newer hardware, especially since support was only recently added. Thanks!

19 minutes ago, Nixon506E said:

I would like to request that the Tehuti drivers that were added in rc16 be removed as it seems to be causing boot issues and they don't seem to be the proper drivers for any of the new Tehuti networking chips. Hopefully removing legacy support due to it causing issues for users of the newer hardware, especially since support was only recently added. Thanks!

 

Maybe we can add a symlink like this:

/etc/modprobe.d/custom  ->  /boot/config/modprobe_custom

This way you can create a file to keep on your flash where you can blacklist that driver if you wanted.  Probably should work.

 

23 minutes ago, limetech said:

 

Maybe we can add a symlink like this:

/etc/modprobe.d/custom  ->  /boot/config/modprobe_custom

This way you can create a file to keep on your flash where you can blacklist that driver if you wanted.  Probably should work.

 

 

That should work, I like the idea of being able to decide which networking drivers are loaded. Thanks.

Ryzen 5 1600 user here... 1 Day 11 Hours uptime so far after changing the C6 State in the go script.... *touch wood*

 

You guys are my lifesavers, I really didn't want to have to offline my array to put a Graphics card in (it's too big and impacts where my drives are) to find why it kept dying.

 

It was kinda fine when I was on rc9 with uptime around a few days, but when I updated to 15, 16 and 17, it would crash close to every 12 hours.

Edited January 1, 20188 yr by Badams

20 hours ago, Nixon506E said:

 

That should work, I like the idea of being able to decide which networking drivers are loaded. Thanks.

 

Actually you can blacklist the driver in the syslinx.cfg file.  Click on Main/Flash and scroll down to the Syslinux configuration box.  Look for the section where you see the line:

  menu default

That is the section being referenced to boot your server.  Next, on the line with "append" add a parameter called "module_blacklist=<module-list>" where <module-list> is a comma-separated list of modules to prevent being loaded by the kernel.

 

For example, if your append line looks like this:

  append initrd=/bzroot,/bzroot-gui

To prevent the tehuti driver being loaded, change to:

  append initrd=/bzroot,/bzroot-gui module_blacklist=tehuti

 

Maybe I am missing something, but I tried searching the forum for it and cannot find it.

 

But it appears that on the Plugins -> Installed Plugins page the Check Updates button is missing. At least on my system it is missing. I am running version 6.4.0_rc18f, upgraded from 6.3.5. The button disappeared after updating. 

 

I get the notifications that there are updates for plugins and recently they would show, but today I got the notifications for CA and Unassigned Devices being updated. I installed them on my 6.3.5 server but on my 6.4 server all of my plugins on the page say unknown for status.

 

How do I get the button back?

 

Update - 1/1/2018 @ 16:31 PDT:  I tried running the script "/usr/local/emhttp/plugins/dynamix.plugin.manager/scripts/plugincheck" manually but it did not work. I ended up rebooting the server and the the updates were there. But still no button on the page for checking updates. Would like to have that back.

Edited January 2, 20188 yr by StanC
Update

12 hours ago, Badams said:

Ryzen 5 1600 user here... 1 Day 11 Hours uptime so far after changing the C6 State in the go script.... *touch wood*

 

You guys are my lifesavers, I really didn't want to have to offline my array to put a Graphics card in (it's too big and impacts where my drives are) to find why it kept dying.

 

It was kinda fine when I was on rc9 with uptime around a few days, but when I updated to 15, 16 and 17, it would crash close to every 12 hours.

 

 

Scratch that, it went down about 6 hours after I posted this...

 

@limetech, did you want me to create a thread for this or tap onto an existing?

8 hours ago, StanC said:

Maybe I am missing something, but I tried searching the forum for it and cannot find it.

 

But it appears that on the Plugins -> Installed Plugins page the Check Updates button is missing. At least on my system it is missing. I am running version 6.4.0_rc18f, upgraded from 6.3.5. The button disappeared after updating. 

 

I get the notifications that there are updates for plugins and recently they would show, but today I got the notifications for CA and Unassigned Devices being updated. I installed them on my 6.3.5 server but on my 6.4 server all of my plugins on the page say unknown for status.

 

How do I get the button back?

 

Update - 1/1/2018 @ 16:31 PDT:  I tried running the script "/usr/local/emhttp/plugins/dynamix.plugin.manager/scripts/plugincheck" manually but it did not work. I ended up rebooting the server and the the updates were there. But still no button on the page for checking updates. Would like to have that back.

 

The button is gone. It checks when you enter the plugins page now (If I remember correctly).

Correct. When plugins page is opened an immediate check is done to ensure the correct state of each plugin.

If the state is "unknown" it means there wasn't a working connection to retrieve the status. Check your network.