-
Notifications
You must be signed in to change notification settings - Fork 170
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
g4 instance works for the first time, but cannot be properly recreated from AMI #16
Comments
These scripts aren't intended to be converted into a AMI so I haven't tested the functionality |
I would test it but the some of the NVidia AMIs are broken as of the 442.19 release today. Looks like they forgot to embed the Amazon security stuff again. I'll certainly try imaging of a vanilla 2019 AMI with the parsec scripts running and see what happens. |
Just note that if you do create an AMI from the Parsec Cloud Prep Tool (this repo) the password will be baked in to the AMI. In addition there are also some static routes that need to be removed before AMI creation, these static routes are region specific, and will completely prevent AMIs copied to other regions from working. If you're building a G4 based AMI and have used GPU updater tool to install the driver, also run |
I am the sole user of the AMI, so the baked in credentials should not matter (right?). If a custom AMI creation is not supported, Is there another option how to launch a spot instance without having to setup everything each time? |
Not right now |
It works fine with a vanilla 2016 server instead of using the Nvidia Gaming. I think you are accidentally using NVidias initialisation script and messing up the drivers after sysprep. Given how bad NVidia are at creating AMI's, I'd suggest not using those images anyway as a general rule. All they have over vanilla is the Nvidia SDK which is downloadable after free registration to the Nvidia developer site. |
Steven, it works for me as well, thank you! It is basically the oposite of this advice:
|
It seems that way but not really. The Grid Virtual workstation are talking
about the official images from the cloud provider with the drivers/Grid
License installed. So Azure, Paperspace & Google all have these when you
deploy a GPU machine. It's just AWS that's a little different. NVidia AMI
is just a third party machine. That message seems specific to Google cloud.
…On Wed, 19 Feb 2020 at 07:14, David ***@***.***> wrote:
Steven, it works for me as well, thank you!
It is basically the oposite of this advice:
Q. My cloud machine is stuck at 1366x768
A. You should delete your machine and use the GRID Virtual Workstation variety provided by your cloud provider (it's a check box on GCP)
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#16>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AGNHJVYVWWB4CNIO6QCJG2TRDTL55ANCNFSM4KWOUSEA>
.
|
FWIW I'm using a combination of an AMI created with this tool along with https://github.com/cloudrig/CloudRIG to launch a ready-made cloud machine using Spot Instance pricing. It works awesome! Thank you @jamesstringerparsec for this work of art - any way to buy you a beer? |
When I run the instance for the first time and use the script exactly as described, all works great. But when I save this instance into an AMI, after recreating it, the resolution is fixed to 1280*1024.
Might be related to #13 with the difference that all works good for the first time.
The text was updated successfully, but these errors were encountered: