@Pan2018iop said:
Hello!
I would like to get an account.
I'm interested on Linux and KVM i want to learn about pci passthough i want to try qemu-kvm
also i'm a thai high school student and now learning about iot and many thing
Here is my SSH key:
ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAID71Dl57AX6v6eTc8pMzwa7bnxj8IXiE5mk6YZX+NmOi pun@DESKTOP-4PDT144
If you had to provide evidence showing that you are a Thai high school student, what evidence could you provide?
What got you interested in PCI passthrough?
What got you interested in iot?
Google found me a guide to non-root PCI passthrough at https://www.evonide.com/non-root-gpu-passthrough-setup/ but I had to use Google cache to read the page. The page has a February 8, 2016 date. Do you have a guide to non-root PCI passthrough which you recommend?
Thanks again for your request and for your interest in MetalVPS!
Best wishes!
Tom
What PCI device are you trying to passthrough?
Is the PCI device available on the MetalVPS machine?
If not, would you be able to provide the device to the datacenter that the machine is physically in, and hire someone to install it?
Q:What PCI device are you trying to passthrough?
A: M.2 PCI SSD or WIFI CARD
Q:Is the PCI device available on the MetalVPS machine?
A: Maybe not
Q:If not, would you be able to provide the device to the datacenter that the machine is physically in, and hire someone to install it?
A: I can't do hire someone to install it.
@Pan2018iop said:
Hello!
I would like to get an account.
I'm interested on Linux and KVM i want to learn about pci passthough i want to try qemu-kvm
also i'm a thai high school student and now learning about iot and many thing
Here is my SSH key:
ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAID71Dl57AX6v6eTc8pMzwa7bnxj8IXiE5mk6YZX+NmOi pun@DESKTOP-4PDT144
If you had to provide evidence showing that you are a Thai high school student, what evidence could you provide?
What got you interested in PCI passthrough?
What got you interested in iot?
Google found me a guide to non-root PCI passthrough at https://www.evonide.com/non-root-gpu-passthrough-setup/ but I had to use Google cache to read the page. The page has a February 8, 2016 date. Do you have a guide to non-root PCI passthrough which you recommend?
Thanks again for your request and for your interest in MetalVPS!
Best wishes!
Tom
Q: What got you interested in iot?
A: I seen many video in youtube that about iot so that why i am interested iot
Q:If you had to provide evidence showing that you are a Thai high school student, what evidence could you provide?
A: My name is Pannarat Wiriyaarritham Student in phitsanulok pittayakom school Grade 9
Q:What got you interested in PCI passthrough?
A: Because i want to do gpu passthrough and i passthrough many thing like wifi card lan card
The old SSH key: Can not be use now
Here the new key
ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIKZo/2iD0bQFLUeyWHO+4PiIXckqWoYkmzVsfuRaBzrW administrator@WIN-HP5SS4NN3PO
Thanks for your message! Welcome to MetalVPS! I will set you up with an account, probably tomorrow, and I will post again here when your account is ready.
Above, I asked:
Google found me a guide to non-root PCI passthrough at https://www.evonide.com/non-root-gpu-passthrough-setup/ but I had to use Google cache to read the page. The page has a February 8, 2016 date. Do you have a guide to non-root PCI passthrough which you recommend?
PCI passthrough might be a little tricky, especially when you are not the only user on the machine and when you do not yet have root privileges. We can talk about pci passthrough some more, and maybe some of the other guys here can teach me more about it.
You are very welcome here! Please do not worry about anything! More soon, probably tomorrow, after I get your account provisioned.
@Exinjh said: Yo, I've got Windows 10 Enterprise LTSB 2016 working in MetalVPS! I'll test performance of my VM (if I allocated so many ram, I can decrease it)
Thanks for your report and for the screenshot! I think everything is okay for now. But when you are done testing, maybe decrease the resources you are using to not more than actually needed. Give the other guys a chance!
Sure! I will decrease my resources!
root@fsn ~ # date; top -n 1 -b | head -n 9
Sun Apr 30 08:31:41 PM UTC 2023
top - 20:31:42 up 1 day, 22:00, 9 users, load average: 14.04, 13.24, 11.27
Tasks: 407 total, 3 running, 404 sleeping, 0 stopped, 0 zombie
%Cpu(s): 30.0 us, 0.0 sy, 0.0 ni, 40.0 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
MiB Mem : 64098.1 total, 394.5 free, 35049.3 used, 29372.1 buff/cache
MiB Swap: 32734.0 total, 26356.5 free, 6377.5 used. 29048.7 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
53929 zeyad 20 0 9786.6m 8.1g 15128 R 800.0 12.9 69:19.03 qemu-sy+
51394 exinjh 20 0 10.0g 7.9g 13800 R 740.0 12.7 23,22 qemu-sy+
root@fsn ~ #
I think it is okay for you guys to run a lot of powerful stuff, but, what exactly are you running? Why? And for how long?
The more energy you devote to answering my questions, the more energy you can use on the node! And for longer.
Best wishes!
Tom
Oh, sorry. I runned benchmarks, and VM in VM (to test the performance of nested virtualization). Again, very sorry for using so many resources. I think it's because I allocated 8 cores for VM, i'll decrease my specs if you will resume my account.
Thanks for your message. Please allow me to ask two questions:
You previously said, "Sure I will decrease my resources!" Please help us understand, "How could we believe you this time when you didn't decrease your resources before?"
@Exinjh said: Yo, I've got Windows 10 Enterprise LTSB 2016 working in MetalVPS! I'll test performance of my VM (if I allocated so many ram, I can decrease it)
Thanks for your report and for the screenshot! I think everything is okay for now. But when you are done testing, maybe decrease the resources you are using to not more than actually needed. Give the other guys a chance!
Sure! I will decrease my resources!
root@fsn ~ # date; top -n 1 -b | head -n 9
Sun Apr 30 08:31:41 PM UTC 2023
top - 20:31:42 up 1 day, 22:00, 9 users, load average: 14.04, 13.24, 11.27
Tasks: 407 total, 3 running, 404 sleeping, 0 stopped, 0 zombie
%Cpu(s): 30.0 us, 0.0 sy, 0.0 ni, 40.0 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
MiB Mem : 64098.1 total, 394.5 free, 35049.3 used, 29372.1 buff/cache
MiB Swap: 32734.0 total, 26356.5 free, 6377.5 used. 29048.7 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
53929 zeyad 20 0 9786.6m 8.1g 15128 R 800.0 12.9 69:19.03 qemu-sy+
51394 exinjh 20 0 10.0g 7.9g 13800 R 740.0 12.7 23,22 qemu-sy+
root@fsn ~ #
I think it is okay for you guys to run a lot of powerful stuff, but, what exactly are you running? Why? And for how long?
The more energy you devote to answering my questions, the more energy you can use on the node! And for longer.
Best wishes!
Tom
Oh, sorry. I runned benchmarks, and VM in VM (to test the performance of nested virtualization). Again, very sorry for using so many resources. I think it's because I allocated 8 cores for VM, i'll decrease my specs if you will resume my account.
Thanks for your message. Please allow me to ask two questions:
You previously said, "Sure I will decrease my resources!" Please help us understand, "How could we believe you this time when you didn't decrease your resources before?"
Thanks again for your message and for your interest in MetalVPS! I look forward to resuming your account!
Best!
Tom
I forgot to decrease my resources. I'm so sorry. I will ask my friend to write something about me if you'll ask me to provide proof that I'm really 17 years old and from France.
@Exinjh said: Yo, I've got Windows 10 Enterprise LTSB 2016 working in MetalVPS! I'll test performance of my VM (if I allocated so many ram, I can decrease it)
Thanks for your report and for the screenshot! I think everything is okay for now. But when you are done testing, maybe decrease the resources you are using to not more than actually needed. Give the other guys a chance!
Sure! I will decrease my resources!
root@fsn ~ # date; top -n 1 -b | head -n 9
Sun Apr 30 08:31:41 PM UTC 2023
top - 20:31:42 up 1 day, 22:00, 9 users, load average: 14.04, 13.24, 11.27
Tasks: 407 total, 3 running, 404 sleeping, 0 stopped, 0 zombie
%Cpu(s): 30.0 us, 0.0 sy, 0.0 ni, 40.0 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
MiB Mem : 64098.1 total, 394.5 free, 35049.3 used, 29372.1 buff/cache
MiB Swap: 32734.0 total, 26356.5 free, 6377.5 used. 29048.7 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
53929 zeyad 20 0 9786.6m 8.1g 15128 R 800.0 12.9 69:19.03 qemu-sy+
51394 exinjh 20 0 10.0g 7.9g 13800 R 740.0 12.7 23,22 qemu-sy+
root@fsn ~ #
I think it is okay for you guys to run a lot of powerful stuff, but, what exactly are you running? Why? And for how long?
The more energy you devote to answering my questions, the more energy you can use on the node! And for longer.
Best wishes!
Tom
Hi Very sorry am using a lot of resources currently qemu vm which is the only thing I do in metalvps has 8 core and 8 GB ram the moment you captured was I probably running a heavy game on Qemu VM and I setted up rdpwrapper so I and my friend can connect
Please tell me I need to lower resoruces
If am running something heavy it wont be for long am very sorry btw
What game was it that you were playing? How many friends?
What do our other neighbors on the server think about allowing you to continue? Have you posted about what you are doing? When you want to do it? How long it will take?
@Exinjh said: Yo, I've got Windows 10 Enterprise LTSB 2016 working in MetalVPS! I'll test performance of my VM (if I allocated so many ram, I can decrease it)
Thanks for your report and for the screenshot! I think everything is okay for now. But when you are done testing, maybe decrease the resources you are using to not more than actually needed. Give the other guys a chance!
Sure! I will decrease my resources!
root@fsn ~ # date; top -n 1 -b | head -n 9
Sun Apr 30 08:31:41 PM UTC 2023
top - 20:31:42 up 1 day, 22:00, 9 users, load average: 14.04, 13.24, 11.27
Tasks: 407 total, 3 running, 404 sleeping, 0 stopped, 0 zombie
%Cpu(s): 30.0 us, 0.0 sy, 0.0 ni, 40.0 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
MiB Mem : 64098.1 total, 394.5 free, 35049.3 used, 29372.1 buff/cache
MiB Swap: 32734.0 total, 26356.5 free, 6377.5 used. 29048.7 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
53929 zeyad 20 0 9786.6m 8.1g 15128 R 800.0 12.9 69:19.03 qemu-sy+
51394 exinjh 20 0 10.0g 7.9g 13800 R 740.0 12.7 23,22 qemu-sy+
root@fsn ~ #
I think it is okay for you guys to run a lot of powerful stuff, but, what exactly are you running? Why? And for how long?
The more energy you devote to answering my questions, the more energy you can use on the node! And for longer.
Best wishes!
Tom
Oh, sorry. I runned benchmarks, and VM in VM (to test the performance of nested virtualization). Again, very sorry for using so many resources. I think it's because I allocated 8 cores for VM, i'll decrease my specs if you will resume my account.
Thanks for your message. Please allow me to ask two questions:
You previously said, "Sure I will decrease my resources!" Please help us understand, "How could we believe you this time when you didn't decrease your resources before?"
Thanks again for your message and for your interest in MetalVPS! I look forward to resuming your account!
Best!
Tom
I forgot to decrease my resources. I'm so sorry. I will ask my friend to write something about me if you'll ask me to provide proof that I'm really 17 years old and from France.
@Exinjh said: Yo, I've got Windows 10 Enterprise LTSB 2016 working in MetalVPS! I'll test performance of my VM (if I allocated so many ram, I can decrease it)
Thanks for your report and for the screenshot! I think everything is okay for now. But when you are done testing, maybe decrease the resources you are using to not more than actually needed. Give the other guys a chance!
Sure! I will decrease my resources!
root@fsn ~ # date; top -n 1 -b | head -n 9
Sun Apr 30 08:31:41 PM UTC 2023
top - 20:31:42 up 1 day, 22:00, 9 users, load average: 14.04, 13.24, 11.27
Tasks: 407 total, 3 running, 404 sleeping, 0 stopped, 0 zombie
%Cpu(s): 30.0 us, 0.0 sy, 0.0 ni, 40.0 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
MiB Mem : 64098.1 total, 394.5 free, 35049.3 used, 29372.1 buff/cache
MiB Swap: 32734.0 total, 26356.5 free, 6377.5 used. 29048.7 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
53929 zeyad 20 0 9786.6m 8.1g 15128 R 800.0 12.9 69:19.03 qemu-sy+
51394 exinjh 20 0 10.0g 7.9g 13800 R 740.0 12.7 23,22 qemu-sy+
root@fsn ~ #
I think it is okay for you guys to run a lot of powerful stuff, but, what exactly are you running? Why? And for how long?
The more energy you devote to answering my questions, the more energy you can use on the node! And for longer.
Best wishes!
Tom
Oh, sorry. I runned benchmarks, and VM in VM (to test the performance of nested virtualization). Again, very sorry for using so many resources. I think it's because I allocated 8 cores for VM, i'll decrease my specs if you will resume my account.
Thanks for your message. Please allow me to ask two questions:
You previously said, "Sure I will decrease my resources!" Please help us understand, "How could we believe you this time when you didn't decrease your resources before?"
Thanks again for your message and for your interest in MetalVPS! I look forward to resuming your account!
Best!
Tom
I forgot to decrease my resources. I'm so sorry. I will ask my friend to write something about me if you'll ask me to provide proof that I'm really 17 years old and from France.
@Exinjh said: Yo, I've got Windows 10 Enterprise LTSB 2016 working in MetalVPS! I'll test performance of my VM (if I allocated so many ram, I can decrease it)
Thanks for your report and for the screenshot! I think everything is okay for now. But when you are done testing, maybe decrease the resources you are using to not more than actually needed. Give the other guys a chance!
Sure! I will decrease my resources!
root@fsn ~ # date; top -n 1 -b | head -n 9
Sun Apr 30 08:31:41 PM UTC 2023
top - 20:31:42 up 1 day, 22:00, 9 users, load average: 14.04, 13.24, 11.27
Tasks: 407 total, 3 running, 404 sleeping, 0 stopped, 0 zombie
%Cpu(s): 30.0 us, 0.0 sy, 0.0 ni, 40.0 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
MiB Mem : 64098.1 total, 394.5 free, 35049.3 used, 29372.1 buff/cache
MiB Swap: 32734.0 total, 26356.5 free, 6377.5 used. 29048.7 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
53929 zeyad 20 0 9786.6m 8.1g 15128 R 800.0 12.9 69:19.03 qemu-sy+
51394 exinjh 20 0 10.0g 7.9g 13800 R 740.0 12.7 23,22 qemu-sy+
root@fsn ~ #
I think it is okay for you guys to run a lot of powerful stuff, but, what exactly are you running? Why? And for how long?
The more energy you devote to answering my questions, the more energy you can use on the node! And for longer.
Best wishes!
Tom
Oh, sorry. I runned benchmarks, and VM in VM (to test the performance of nested virtualization). Again, very sorry for using so many resources. I think it's because I allocated 8 cores for VM, i'll decrease my specs if you will resume my account.
Thanks for your message. Please allow me to ask two questions:
You previously said, "Sure I will decrease my resources!" Please help us understand, "How could we believe you this time when you didn't decrease your resources before?"
Thanks again for your message and for your interest in MetalVPS! I look forward to resuming your account!
Best!
Tom
I forgot to decrease my resources. I'm so sorry. I will ask my friend to write something about me if you'll ask me to provide proof that I'm really 17 years old and from France.
root@fsn /home/pan2018iop # ls -alR
.:
total 28
drwx------ 3 pan2018iop pan2018iop 4096 May 1 06:54 .
drwxr-xr-x 31 root root 4096 May 1 06:51 ..
-rw-r--r-- 1 pan2018iop pan2018iop 220 May 1 06:51 .bash_logout
-rw-r--r-- 1 pan2018iop pan2018iop 3526 May 1 06:51 .bashrc
-rw-r--r-- 1 pan2018iop pan2018iop 0 May 1 06:51 .cloud-locale-test.skip
-rw------- 1 pan2018iop pan2018iop 17 May 1 06:54 password
-rw-r--r-- 1 pan2018iop pan2018iop 807 May 1 06:51 .profile
drwx------ 2 pan2018iop pan2018iop 4096 May 1 06:55 .ssh
./.ssh:
total 12
drwx------ 2 pan2018iop pan2018iop 4096 May 1 06:55 .
drwx------ 3 pan2018iop pan2018iop 4096 May 1 06:54 ..
-rw------- 1 pan2018iop pan2018iop 111 May 1 06:55 authorized_keys
root@fsn /home/pan2018iop # cat .ssh/authorized_keys
ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIKZo/2iD0bQFLUeyWHO+4PiIXckqWoYkmzVsfuRaBzrW administrator@WIN-HP5SS4NN3PO
root@fsn /home/pan2018iop #
Here are some basics that you need to keep in mind:
Password access has been disabled. You need your key to log in.
Your password is in a mode 600 file in your home directory.
Your password might be needed later for sudo. But you have not yet been added to the sudo group.
Feel free to change your password.
You have been added to the kvm group.
If you want to make a KVM VPS it is recommended to start with slirp following this tutorial.
Alternatives include setting up proot as described earlier in this thread.
Additional node configuration might be needed for LXC and non-slirp qemu.
Please post your questions and progress reports and tutorials here in this thread. It benefits everyone to help answer everyone else's questions and to learn from everyone else's progress reports and tutorials. So, please post here instead of sending me PMs (PMs are okay if privacy really is necessary).
Have fun! Thanks for asking for a MetalVPS account!
@Exinjh said: Yo, I've got Windows 10 Enterprise LTSB 2016 working in MetalVPS! I'll test performance of my VM (if I allocated so many ram, I can decrease it)
Thanks for your report and for the screenshot! I think everything is okay for now. But when you are done testing, maybe decrease the resources you are using to not more than actually needed. Give the other guys a chance!
Sure! I will decrease my resources!
root@fsn ~ # date; top -n 1 -b | head -n 9
Sun Apr 30 08:31:41 PM UTC 2023
top - 20:31:42 up 1 day, 22:00, 9 users, load average: 14.04, 13.24, 11.27
Tasks: 407 total, 3 running, 404 sleeping, 0 stopped, 0 zombie
%Cpu(s): 30.0 us, 0.0 sy, 0.0 ni, 40.0 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
MiB Mem : 64098.1 total, 394.5 free, 35049.3 used, 29372.1 buff/cache
MiB Swap: 32734.0 total, 26356.5 free, 6377.5 used. 29048.7 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
53929 zeyad 20 0 9786.6m 8.1g 15128 R 800.0 12.9 69:19.03 qemu-sy+
51394 exinjh 20 0 10.0g 7.9g 13800 R 740.0 12.7 23,22 qemu-sy+
root@fsn ~ #
I think it is okay for you guys to run a lot of powerful stuff, but, what exactly are you running? Why? And for how long?
The more energy you devote to answering my questions, the more energy you can use on the node! And for longer.
Best wishes!
Tom
Oh, sorry. I runned benchmarks, and VM in VM (to test the performance of nested virtualization). Again, very sorry for using so many resources. I think it's because I allocated 8 cores for VM, i'll decrease my specs if you will resume my account.
Thanks for your message. Please allow me to ask two questions:
You previously said, "Sure I will decrease my resources!" Please help us understand, "How could we believe you this time when you didn't decrease your resources before?"
Thanks again for your message and for your interest in MetalVPS! I look forward to resuming your account!
Best!
Tom
I forgot to decrease my resources. I'm so sorry. I will ask my friend to write something about me if you'll ask me to provide proof that I'm really 17 years old and from France.
@Pan2018iop said:
Hello!
I would like to get an account.
I'm interested on Linux and KVM i want to learn about pci passthough i want to try qemu-kvm
also i'm a thai high school student and now learning about iot and many thing
Here is my SSH key:
ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAID71Dl57AX6v6eTc8pMzwa7bnxj8IXiE5mk6YZX+NmOi pun@DESKTOP-4PDT144
If you had to provide evidence showing that you are a Thai high school student, what evidence could you provide?
What got you interested in PCI passthrough?
What got you interested in iot?
Google found me a guide to non-root PCI passthrough at https://www.evonide.com/non-root-gpu-passthrough-setup/ but I had to use Google cache to read the page. The page has a February 8, 2016 date. Do you have a guide to non-root PCI passthrough which you recommend?
Thanks again for your request and for your interest in MetalVPS!
Best wishes!
Tom
Q: What got you interested in iot?
A: I seen many video in youtube that about iot so that why i am interested iot
Q:If you had to provide evidence showing that you are a Thai high school student, what evidence could you provide?
A: My name is Pannarat Wiriyaarritham Student in phitsanulok pittayakom school Grade 9
Q:What got you interested in PCI passthrough?
A: Because i want to do gpu passthrough and i passthrough many thing like wifi card lan card
The old SSH key: Can not be use now
Here the new key
ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIKZo/2iD0bQFLUeyWHO+4PiIXckqWoYkmzVsfuRaBzrW administrator@WIN-HP5SS4NN3PO
Thanks for your message! Welcome to MetalVPS! I will set you up with an account, probably tomorrow, and I will post again here when your account is ready.
Above, I asked:
Google found me a guide to non-root PCI passthrough at https://www.evonide.com/non-root-gpu-passthrough-setup/ but I had to use Google cache to read the page. The page has a February 8, 2016 date. Do you have a guide to non-root PCI passthrough which you recommend?
PCI passthrough might be a little tricky, especially when you are not the only user on the machine and when you do not yet have root privileges. We can talk about pci passthrough some more, and maybe some of the other guys here can teach me more about it.
You are very welcome here! Please do not worry about anything! More soon, probably tomorrow, after I get your account provisioned.
Best wishes and thanks again!
Tom
Based on his response he is not ready for PCI passthrough on MetalVPS.
IoT? Maybe, but "IoT" is a topic too generic.
I have a question about load averages on this i9-13900 server: What load average means that additional processor intensive tasks will receive significantly lower performance?
# of Performance-cores 8
# of Efficient-cores 16
Total Threads 32
With 8 performance-cores it might seem like a load average of 8 or more could suggest that the performance-core capacity of the server could be close to fully utilized.
Here is what I saw a moment ago on the server:
root@fsn ~ # date; top -n 1 -b | head -n 15
Mon May 1 05:50:56 PM UTC 2023
top - 17:50:56 up 2 days, 19:19, 10 users, load average: 7.56, 6.36, 4.98
Tasks: 410 total, 1 running, 409 sleeping, 0 stopped, 0 zombie
%Cpu(s): 12.5 us, 6.2 sy, 0.0 ni, 68.8 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
MiB Mem : 64098.1 total, 541.9 free, 47824.8 used, 16449.2 buff/cache
MiB Swap: 32734.0 total, 29398.6 free, 3335.4 used. 16273.2 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
11502 redacted 20 0 4562936 2.8g 5196 S 240.0 4.5 22,42 qemu-sy+
11028 redacted 20 0 3229252 1.9g 5560 S 186.7 3.0 20,00 qemu-sy+
66404 redacted 20 0 12.3g 8.0g 11004 S 120.0 12.8 294:56.03 qemu-sy+
68671 redacyed 20 0 25.2g 23.7g 12372 S 73.3 37.9 24:20.18 qemu-sy+
68262 redacted 20 0 9484640 8.0g 11924 S 40.0 12.8 129:10.33 qemu-sy+
So, when the load average, which was 7.56, reaches or rises above 8, would it be correct to assume that the CPU might be low on performance-core capacity, and that additional processor intensive tasks will receive significantly lower performance?
@Not_Oles said: @iamvinh123 Installing in RAM is fun! I just got another server with 128GB of RAM, so more memory based installs might be possible on the other server. It doesn't have anyone using it yet besides me. It's an i9-9900K, so the processor is maybe only half as fast as this i9-13900 server for some tasks, but it has double the RAM. If you want to try it, let me know.
@dinopotato Have you tried installing Linux yet? You don't have to blow away your Windows install. If you want to try Linux, just make a second VM.
Actually i have one Ubuntu too, i also want to use Kali Linux and Arch Linux but i didnt installed it yet
i also try virtuallation in my vm but not success because when i restart it windows update blow my vm up ^^
I success at this expirement again, i download a file from my vm use a tool call "IDM (aka Internet download manager)" to download a file which is super slow because the host is potato then i uploaded to MetalVPS, I can even download it to my main PC too
@iamvinh123 Installing in RAM is fun! I just got another server with 128GB of RAM, so more memory based installs might be possible on the other server. It doesn't have anyone using it yet besides me. It's an i9-9900K, so the processor is maybe only half as fast as this i9-13900 server for some tasks, but it has double the RAM. If you want to try it, let me know.
@Not_Oles said:
I have a question about load averages on this i9-13900 server: What load average means that additional processor intensive tasks will receive significantly lower performance?
So, the thing to remember is that the load is the number of threads/processes that are ready to execute immediately.
With 8 performance-cores it might seem like a load average of 8 or more could suggest that the performance-core capacity of the server could be close to fully utilized.
So, when the load average, which was 7.56, reaches or rises above 8, would it be correct to assume that the CPU might be low on performance-core capacity, and that additional processor intensive tasks will receive significantly lower performance?
I don't think you can particularly reason that way, and I think there's possibly still more work needed in the kernel to do what you think is happening...
So, the current scheduler (CFQ) tries to maintain fairness by working out the CPU time each process has had recently and uses that to adjust the priorities if there are more threads waiting than cores. Additionally, there's a whole load of other stuff that tries to keep running tasks running longer on the same core, to avoid context switches and wasting the contents of the cache (which is pretty much useless if another process is run instead).
If the load is only 8 and you have 32 logical cores/threads, then the CFQ will almost certainly try to keep running tasks on the core they are already on, and put previously sleeping tasks on a new core when they wake up. So, there's not really any guarantee that things will even be running on the performance cores anyway. This is mitigated a bit by the scheduler trying to put newly woken threads on the lowest number core, so usually the lower number cores will be busy and the higher ones not so much. This is also why an n-core, 2n-thread CPU has logical cores numbered in pairs 0..(n-1) and n..(2n-1), because that means that if only n logical cores are needed, they're all on their own physical core rather than using 2 threads that share resources on a single core.
I've no experience on what happens with performance / efficiency cores under Linux and from what I can tell, most of the work on schedulers was done before they became common, so it's possible that Linux just considers them all equal and you get what you get. If the cores are enumerated performance first, then efficiency, then what you surmised might be the case.
I know under MacOS, it's quite complicated, and under low power use, the OS tries to actively move tasks to the slower efficiency cores so that it can power off the performance cores, but the balance shifts when those cores become busy and it'll then try to load up the performance cores and shut down the efficiency cores if it can.
Originally, I was going to ask for an account on the machine to try some experiments to profile this kind of thing, but I know I don't actually have enough time to do it, so I didn't ask.
One thing I will suggest though is pinning some people's VMs to certain cores only, so if you had 8 users, you could maybe allocate each a performance core that nobody else can touch and then let them allocate the rest of their cores from the pool of the other cores (I'm not actually sure whether it's efficiency or performance that has 2 threads per core on these chips, so maybe this suggestion should be flipped round).
Please let me add that I don't even know yet whether the new Thread Director tech is enabled in the kernel we are running at the moment. I'm just beginning to learn about this. Seems like it might be good to enable it.
Thanks, as always, to @ralf for interesting and helpful comments!
You guys are welcome to enjoy the i9-9900K server with 128 GB RAM for a little while. How long? A few days to a few weeks. At least for now, you two guys are the only people using the server.
On the i9-9900K you both are members of the kvm group and also the sudo group. The server is running Debian sid. The setup is very similar to the i9-13900. You should feel free to install whatever additional software you want as long as it is "white hat."
Please don't worry about breaking the server. It's easy to reinstall. Reboots sometimes need a power interrupt. If you reboot the server and it doesn't come back up, just let me know.
If you want to try LXC or giving each KVM its own IPv4 port forwards or IPv6 addresses, the i9-9900K needs additional configuration. We can do the additional configuration if you want.
I prefer open source software, so, if you install Windows or other closed source software for a brief test, please wipe it as soon as you are finished testing.
Your accounts on the i9-13900 are still going, so you are welcome to continue using the i9-13900 throughout.
Please continue to post here about what you are doing and about any questions. Hope you have fun on the i9-9900K! And on the i9-13900!
@yqua I added your key to the root account. If you want a user account, please feel free to make it.
Other guys, do you want to be on the i9-9900K? I am not sure what I want to do about additional users, but we will see what happens if anybody asks.
You guys are welcome to enjoy the i9-9900K server with 128 GB RAM for a little while. How long? A few days to a few weeks. At least for now, you two guys are the only people using the server.
On the i9-9900K you both are members of the kvm group and also the sudo group. The server is running Debian sid. The setup is very similar to the i9-13900. You should feel free to install whatever additional software you want as long as it is "white hat."
Please don't worry about breaking the server. It's easy to reinstall. Reboots sometimes need a power interrupt. If you reboot the server and it doesn't come back up, just let me know.
If you want to try LXC or giving each KVM its own IPv4 port forwards or IPv6 addresses, the i9-9900K needs additional configuration. We can do the additional configuration if you want.
I prefer open source software, so, if you install Windows or other closed source software for a brief test, please wipe it as soon as you are finished testing.
Your accounts on the i9-13900 are still going, so you are welcome to continue using the i9-13900 throughout.
Please continue to post here about what you are doing and about any questions. Hope you have fun on the i9-9900K! And on the i9-13900!
@yqua I added your key to the root account. If you want a user account, please feel free to make it.
Other guys, do you want to be on the i9-9900K? I am not sure what I want to do about additional users, but we will see what happens if anybody asks.
Have a good evening, guys! Best wishes!
Thanks!
Edit: The sudo command works, I'll install the OVMF package for EFI in QEMU VMs
Comments
Q:What PCI device are you trying to passthrough?
A: M.2 PCI SSD or WIFI CARD
Q:Is the PCI device available on the MetalVPS machine?
A: Maybe not
Q:If not, would you be able to provide the device to the datacenter that the machine is physically in, and hire someone to install it?
A: I can't do hire someone to install it.
Hi @Pan2018iop!
Thanks for your message! Welcome to MetalVPS! I will set you up with an account, probably tomorrow, and I will post again here when your account is ready.
Above, I asked:
PCI passthrough might be a little tricky, especially when you are not the only user on the machine and when you do not yet have root privileges. We can talk about pci passthrough some more, and maybe some of the other guys here can teach me more about it.
You are very welcome here! Please do not worry about anything! More soon, probably tomorrow, after I get your account provisioned.
Best wishes and thanks again!
Tom
I hope everyone gets the servers they want!
Alright, thanks
Hi @Exinjh!
Thanks for your message. Please allow me to ask two questions:
You previously said, "Sure I will decrease my resources!" Please help us understand, "How could we believe you this time when you didn't decrease your resources before?"
At https://lowendspirit.com/discussion/comment/137147/#Comment_137147 you said you were 17 years old and from France. If we asked you to provide proof that you are 17 and from France, what evidence could you provide?
Thanks again for your message and for your interest in MetalVPS! I look forward to resuming your account!
Best!
Tom
I hope everyone gets the servers they want!
I forgot to decrease my resources. I'm so sorry. I will ask my friend to write something about me if you'll ask me to provide proof that I'm really 17 years old and from France.
Hi @NotRealZeyad!
Thanks for your message!
What game was it that you were playing? How many friends?
What do our other neighbors on the server think about allowing you to continue? Have you posted about what you are doing? When you want to do it? How long it will take?
Best wishes!
Tom
I hope everyone gets the servers they want!
@Not_Oles
Hi again @Exinjh!
Thanks for your message! What do other neighbors from our community think about restoring your account?
Best wishes!
Tom
I hope everyone gets the servers they want!
I don't know. Should I ask them?
Hi @Pan2018iop!
I made your account. Hope you have fun on the server! Again, welcome to MetalVPS!
To log in, please try something like
ssh [email protected] -p 42365Here you can check my work:
Here are some basics that you need to keep in mind:
Have fun! Thanks for asking for a MetalVPS account!
Best!
Tom
I hope everyone gets the servers they want!
Please do.
I hope everyone gets the servers they want!
@Not_Oles Can you unsuspend my vm?
Guys, who reading this. Please, answer my question: Should Not_Oles restore my account? Again, sorry for abusing the VPS.
Based on his response he is not ready for PCI passthrough on MetalVPS.
IoT? Maybe, but "IoT" is a topic too generic.
The all seeing eye sees everything...
@Not_Oles can you have a look at this link
i gonna do iGPU
https://wiki.archlinux.org/title/Intel_GVT-g
So what makes you think MetalVPS is a good spot for it?
The all seeing eye sees everything...
Honestly, "PCI passthrough" seems like a ChatGPT response to me.
As far as I can tell from your linked wiki, that wouldn't work on this server anyway.
"Intel GVT-g currently only works with Intel Broadwell (5th gen) to Comet Lake (10th gen)".
But this server has a 13th generation Intel CPU.
Can you elaborate a litte bit more on how you are going to get it to work on this server?
it high spec> @Yqua said:
i might find a way to do it so
i have to learn more about it
Hey, guys.
Answering your question, it sounds like suspending your account was the correct decision.
Oh the hilarity! What makes you think this machine even has a webcam?
Hi Guys!
I have a question about load averages on this i9-13900 server: What load average means that additional processor intensive tasks will receive significantly lower performance?
Let's keep in mind that Linux load averages include disk activity.
From the Intel i9-13900 processor specs:
With 8 performance-cores it might seem like a load average of 8 or more could suggest that the performance-core capacity of the server could be close to fully utilized.
Here is what I saw a moment ago on the server:
So, when the load average, which was 7.56, reaches or rises above 8, would it be correct to assume that the CPU might be low on performance-core capacity, and that additional processor intensive tasks will receive significantly lower performance?
Thanks for your help! Best wishes!
Tom
I hope everyone gets the servers they want!
It won't work.
The all seeing eye sees everything...
I just installed the
sysstatpackage which containsmpstat. Here's an example ofmpstatoutput:Summary of all 32 processors:
Summary plus each processor individually:
I hope everyone gets the servers they want!
Actually i have one Ubuntu too, i also want to use Kali Linux and Arch Linux but i didnt installed it yet
i also try virtuallation in my vm but not success because when i restart it windows update blow my vm up ^^
I success at this expirement again, i download a file from my vm use a tool call "IDM (aka Internet download manager)" to download a file which is super slow because the host is potato then i uploaded to MetalVPS, I can even download it to my main PC too
Can i try it also?
So, the thing to remember is that the load is the number of threads/processes that are ready to execute immediately.
I don't think you can particularly reason that way, and I think there's possibly still more work needed in the kernel to do what you think is happening...
So, the current scheduler (CFQ) tries to maintain fairness by working out the CPU time each process has had recently and uses that to adjust the priorities if there are more threads waiting than cores. Additionally, there's a whole load of other stuff that tries to keep running tasks running longer on the same core, to avoid context switches and wasting the contents of the cache (which is pretty much useless if another process is run instead).
If the load is only 8 and you have 32 logical cores/threads, then the CFQ will almost certainly try to keep running tasks on the core they are already on, and put previously sleeping tasks on a new core when they wake up. So, there's not really any guarantee that things will even be running on the performance cores anyway. This is mitigated a bit by the scheduler trying to put newly woken threads on the lowest number core, so usually the lower number cores will be busy and the higher ones not so much. This is also why an n-core, 2n-thread CPU has logical cores numbered in pairs 0..(n-1) and n..(2n-1), because that means that if only n logical cores are needed, they're all on their own physical core rather than using 2 threads that share resources on a single core.
I've no experience on what happens with performance / efficiency cores under Linux and from what I can tell, most of the work on schedulers was done before they became common, so it's possible that Linux just considers them all equal and you get what you get. If the cores are enumerated performance first, then efficiency, then what you surmised might be the case.
I know under MacOS, it's quite complicated, and under low power use, the OS tries to actively move tasks to the slower efficiency cores so that it can power off the performance cores, but the balance shifts when those cores become busy and it'll then try to load up the performance cores and shut down the efficiency cores if it can.
Originally, I was going to ask for an account on the machine to try some experiments to profile this kind of thing, but I know I don't actually have enough time to do it, so I didn't ask.
One thing I will suggest though is pinning some people's VMs to certain cores only, so if you had 8 users, you could maybe allocate each a performance core that nobody else can touch and then let them allocate the rest of their cores from the pool of the other cores (I'm not actually sure whether it's efficiency or performance that has 2 threads per core on these chips, so maybe this suggestion should be flipped round).
Hi @dinopotato!
Looks like you've been having a lot of fun trying various Linux distros! 🍰
Yeah, sure, you can try the other server along with @iamvinh123. I will let you know when it is ready.
Best wishes!
Tom
I hope everyone gets the servers they want!
Same with me!
Intel has "Thread Director." There is a description of Thread Director in How 13th Gen Intel® Core™ Processors Work.
Intel submitted Linux kernel patches for "Thread Director." The initial set of Thread Director patches might be here.
For NetBSD, I seem to remember hearing about a change in NetBSD 10 described as "Make the scheduler topology aware. It understands and makes good use of HyperThreading/SMT, spreads the load evenly across different physical CPU packages, and can effectively run asymmetric systems with slow and fast CPUs [emphasis added] [ad 20200113]" IIRC, this "asymmetric slow and fast CPUs" change also is mentioned in NetBSD 10: 3 Years in the Making - Nia Alarie - EuroBSDcon 2022.
Please let me add that I don't even know yet whether the new Thread Director tech is enabled in the kernel we are running at the moment. I'm just beginning to learn about this. Seems like it might be good to enable it.
Thanks, as always, to @ralf for interesting and helpful comments!
I hope everyone gets the servers they want!
@iamvinh123
Please try:
ssh [email protected] -p 42365When you get in, please also try:
sudo whoami@dinopotato
Please try:
ssh [email protected] -p 42365When you get in, please also try:
sudo whoami@iamvinh123 @dinopotato
You guys are welcome to enjoy the i9-9900K server with 128 GB RAM for a little while. How long? A few days to a few weeks. At least for now, you two guys are the only people using the server.
On the i9-9900K you both are members of the
kvmgroup and also thesudogroup. The server is running Debian sid. The setup is very similar to the i9-13900. You should feel free to install whatever additional software you want as long as it is "white hat."Please don't worry about breaking the server. It's easy to reinstall. Reboots sometimes need a power interrupt. If you reboot the server and it doesn't come back up, just let me know.
If you want to try LXC or giving each KVM its own IPv4 port forwards or IPv6 addresses, the i9-9900K needs additional configuration. We can do the additional configuration if you want.
I prefer open source software, so, if you install Windows or other closed source software for a brief test, please wipe it as soon as you are finished testing.
Your accounts on the i9-13900 are still going, so you are welcome to continue using the i9-13900 throughout.
Please continue to post here about what you are doing and about any questions. Hope you have fun on the i9-9900K!
And on the i9-13900!
@yqua I added your key to the root account. If you want a user account, please feel free to make it.
Other guys, do you want to be on the i9-9900K? I am not sure what I want to do about additional users, but we will see what happens if anybody asks.
Have a good evening, guys! Best wishes!
I hope everyone gets the servers they want!
Thanks!
Edit: The sudo command works, I'll install the OVMF package for EFI in QEMU VMs