We tried with many regions
We started with UAE north abd UAE central . Then north europe . Wedt europe. France central . West us 2 . West us 3 . East us . Australia . I think these are all the regions we tried .
We tried with many regions
We started with UAE north abd UAE central . Then north europe . Wedt europe. France central . West us 2 . West us 3 . East us . Australia . I think these are all the regions we tried .
Hi @infoatcamouflage1, I just deployed to East US for the both the shared Resource Group and streaming region group and was successful. Iâll follow up with you offline to see if we can look at what settings were being used to deploy it, and we can reproduce it together.
The problem is that customer regions should be UAE central and UAE north . And hoping in the future to have more regions in the middle east .
Hi, whatâs the best way to customize the image used in this Azure Markerplace solution?
Iâve created the Unreal Pixel Streaming app, but when I connect to the traffic manager endpoint I just get; âAll Cirrus servers are in use. Retrying inâŚâ If I look at the logs on the MatchMaker VM it states; Current VMSS count: 1 - Current Cirrus Servers Connected: 2 - Current Cirrus Servers with clients: 0. Iâm really stuck as to what is happening. Any thoughts?
Is this a problem with UE5? UE5 doesnât create a WIndowsNoEditor folder for packaging with PixelStreaming. It appears this market place assumes UE4 and is therefore not finding the EXE??
Apologies on the delayed response here. Unfortunately the machine image that is used for the marketplace solution cannot be customized. The main opportunity for customization is editing the pixel streaming servers to have your own unique web page and other features around that piece.
This almost always means that the UE app was never successfully launched as part of the automation. After uploading the zip of your project there is an input for âUE4 App Nameâ. This field must 100% match the name of your executable (without the â.exeâ). If it doesnât match, it will not launch the right file.
Similarly, the folder structure of your zip file is very important. We have already accounted for the differences in âWindowsâ vs âWindowsNoEditorâ, but the script is expecting this exact configuration of subfolders: Project.zip\Windows\Project.exe . If there is no Windows folder or an extra folder, it may not work. In a future release we intend to avoid the human error involved in this portion.
There is one more edge case where your app is being launched but immediately crashes due to some resource constraint that was not accounted for. If you can remote into an instance and use the app normally, that would rule this out.
1 Like
Thanks, Iâll double check the zip. I manually copied the contents of the zips windows folder into the VMs WindowsNoEditor folder and that âfixedâ the problem on that VM.
Hi . I have confirmed the zip is setup as you suggested. I.e. If I open the zip then the address is MyProject.zip->Windows->MyProject.exe.
I just tested the marketplace deployment with the UE5 First Person project with the Windows subfolder just as you have it. I canât reproduce this error.
You might want to retry the deployment and double check that the âApp Nameâ didnât have a typo in it.
Iâve used the dashboard to upload a zip but changed the folder to NoEditor and it works ??!
We will take a look at it⌠That definitely shouldnât be the case. Thanks for reporting back.
It might be a symptom of something else, e.g. the exe is 4gb and the original resulting unreal_1 folder has barely anything in it. Also PS, the instructions regarding configuration of the dashboard with Azure Directory are out of date. Need to add the token checkboxes.
Gâday. Iâm just testing things out but continue to get deployment failed. For the test I have packaged the UE 5.1 first person template. Is it even possible to deploy using a free trial Azure subscription ?
Hi @jaspos, can you provide any details for what is failing on the deployment? Iâm assuming that you arenât setup for the GPU quota perhaps? Unfortunately there are no free trial subscriptions for deploying GPU solutions in the cloud, and you will need to have at least a Pay as You Go account to get that GPU quota. If/when you do get a Pay as You Go account, request your GPU quota via the portal, and then please reach out with a ticket number if you donât get automatically approved so we can help expedite that.
Found error info that suggests the quota thing. This is part of it.
âOperation could not be completed as it results in exceeding approved Total Regional Cores quota. Additional details - Deployment Model: Resource Manager, Location: southeastasia, Current Limit: 4, Current Usage: 0, Additional Required: 16, (Minimum) New Limit Required: 16. Submit a request for Quota increaseâ
Thanks for the reply.
Canât seem to increase Total Regional vCPUs quota for any location. I have a max of 10 in all locations and it would seem I need this to be 16. Located in Australia Iâve tried the 3 AU locations and then a few random locations but all quota increase requests fail. Is this again something to do with the pay as you go subscription ? I confess to stumbling around in the dark a bit here but all I was hoping to do was upload a build, test the streaming performance and connection , and then delete the deployment and resources.
WellâŚComing back to it, even though getting an Unable to increase message last time, I now have 16 Total Regional vCPUs showing in some of those locations. Now a new error âOperation could not be completed as it results in exceeding approved standardNVSv3Family Cores quota. Additional details - Deployment Model: Resource Manager, Location: southeastasia, Current Limit: 0, Current Usage: 0, Additional Required: 12, (Minimum) New Limit Required: 12.â
Tried increasing the standardNVSv3Family quotas from 0 to 12 but getting the Unable to increase message again and this time the values remain at 0. Tried a new deployment just to check but same failure and error.