Skip to content

Commit 9e4ff28

Browse files
kraigbkraigb
authored
Merge pull request #5 from getroyer/getroyer/DSVMFileSystem
Adding a new page for data science virtual machines.
2 parents a49bf8d + 959f229 commit 9e4ff28

File tree

2 files changed

+67
-26
lines changed

2 files changed

+67
-26
lines changed

‎articles/notebooks/configure-manage-azure-notebooks-projects.md

Lines changed: 2 additions & 26 deletions
Original file line numberDiff line numberDiff line change
@@ -33,32 +33,8 @@ Azure Notebooks starts the underlying virtual machine whenever you run a noteboo
3333

3434
## Compute tier
3535

36-
The **Run** drop-down list on the project dashboard is where you select the compute tier on which the project runs. By default, projects run on the **Free Compute** tier, which is limited to 4GB of memory and 1GB of data to prevent abuse:
37-
38-
![Compute tier drop-down list on the project dashboard](media/project-compute-tier-list.png)
39-
40-
You can bypass these limitations by using a different virtual machine that you've provisioned in an Azure subscription. You must install and run JupyterHub on that virtual machine. The [Data Science Virtual Machine (DSVM) for Linux (Ubuntu)](/azure/machine-learning/data-science-virtual-machine) image is a good choice because it's supported, it includes a compatible JupyterHub by default, and it's automatically discovered by Azure Notebooks.
41-
42-
Once you have a suitably configured Azure virtual machine, select the **Direct Compute** option in the drop-down list, which prompts you for a name (to show in the list), the VM's IP address and port (typically 8000, the default port to which JupyterHub listens), and the VM credentials:
43-
44-
![Prompt to collect server information for the Direct Compute option](media/project-compute-tier-direct.png)
45-
46-
If the following conditions are true, the drop-down list also shows DSVM instances. (If any of these conditions aren't met, you can still connect to the DSVM using the Direct Compute option and enter the values obtained from the Azure portal.)
47-
48-
- You're signed into Azure Notebooks with an account that uses Azure Active Directory (AAD), such as a company account.
49-
- Your account is connected to an Azure subscription.
50-
- You have one or more virtual machines in that subscription, with at least Reader access, that use the Data Science Virtual Machine for Linux (Ubuntu) image.)
51-
52-
![Data Science Virtual Machine instances in the drop-down list on the project dashboard](media/project-compute-tier-dsvm.png)
53-
54-
When you select a DSVM instance, Azure Notebooks may prompt you for the specific machine credentials used when you created the VM.
55-
56-
To create a new DSVM instance, follow the instructions on [Create an Ubuntu Data Science VM](/azure/machine-learning/data-science-virtual-machine/dsvm-ubuntu-intro). Use the **Data Science Virtual Machine for Linux (Ubuntu)** image if you want the DSVM to appear in the drop-down list in Azure Notebooks.
57-
58-
> [!NOTE]
59-
> When a project is run on a VM, the files are mounted on the root directory of the Jupyter server (the directory shown in JupyterHub), replacing the default Azure Notebooks files. When you shut down the VM using the **Shutdown** button on the notebook UI, Azure Notebooks restores the default files.
60-
>
61-
> ![Shutdown button in Azure Notebooks](media/shutdown.png)
36+
By default, projects run on the Free Compute tier, which is limited to 4GB of memory and 1GB of data to prevent abuse. You can bypass these limitations by using a different virtual machine that you've provisioned in an Azure subscription. Connecting to a DSVM can increase your compute power.
37+
See [how to use Data Science Virtual Machines](use-data-science-virtual-machine.md).
6238

6339
## Edit project metadata
6440

‎articles/notebooks/use-data-science-virtual-machine.md

Lines changed: 65 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,65 @@
1+
---
2+
title: Use Data Science Virtual Machines
3+
description: Why use a Data Science Virutal Machine (DSVM). How to connect to a DSVM, and how to access a DSVM from Azure Notebooks.
4+
services: app-service
5+
documentationcenter: ''
6+
author: getroyer
7+
manager: andneil
8+
9+
ms.assetid: 0ccc2529-e17f-4221-b7c7-9496d6a731cc
10+
ms.service: azure-notebooks
11+
ms.workload: na
12+
ms.tgt_pltfrm: na
13+
ms.devlang: na
14+
ms.topic: article
15+
ms.date: 05/08/2019
16+
ms.author: getroyer
17+
---
18+
19+
# Using Data Science Virtual Machines
20+
21+
## Why connect to a Data Science Virtual Machine (DSVM)
22+
23+
By default, projects run on the **Free Compute** tier, which is limited to 4GB of memory and 1GB of data to prevent abuse. You can bypass these limitations by using a different virtual machine that you've provisioned in an Azure subscription. Connecting to a [DSVM](https://azure.microsoft.com/en-us/services/virtual-machines/data-science-virtual-machines/) can increase your compute power.
24+
25+
## How to create a DSVM
26+
27+
|Operating System|Azure Notebooks|
28+
|:---:|:---:|
29+
|Linux Ubuntu| Supported|
30+
|Windows 2016| Not Supported|
31+
|Windows 2012| Not Supported|
32+
|Linux CentOS| Not Supported|
33+
34+
To create a new DSVM instance, follow the instructions on [Create an Ubuntu Data Science VM](/azure/machine-learning/data-science-virtual-machine/dsvm-ubuntu-intro). Use the **Data Science Virtual Machine for Linux (Ubuntu)** image if you want the DSVM to appear in the drop-down list in Azure Notebooks.
35+
36+
## Ways to connect to a DSVM
37+
38+
The **Run** drop-down list on the project dashboard is where you select the compute tier on which the project runs.
39+
40+
![Compute tier drop-down list on the project dashboard](media/project-compute-tier-list.png)
41+
42+
### Discovery
43+
If the following conditions are true, the drop-down list also shows DSVM instances. (If any of these conditions aren't met, you can still connect to the DSVM using the Direct Compute option and enter the values obtained from the Azure portal.)
44+
45+
- You're signed into Azure Notebooks with an account that uses Azure Active Directory (AAD), such as a company account.
46+
- Your account is connected to an Azure subscription.
47+
- You have one or more virtual machines in that subscription, with at least Reader access, that use the Data Science Virtual Machine for Linux (Ubuntu) image.)
48+
49+
![Data Science Virtual Machine instances in the drop-down list on the project dashboard](media/project-compute-tier-dsvm.png)
50+
51+
When you select a DSVM instance, Azure Notebooks may prompt you for the specific machine credentials used when you created the VM.
52+
53+
### Direct Connect
54+
Once you have a suitably configured Azure virtual machine, select the **Direct Compute** option in the drop-down list, which prompts you for a name (to show in the list), the VM's IP address and port (typically 8000, the default port to which JupyterHub listens), and the VM credentials:
55+
56+
![Prompt to collect server information for the Direct Compute option](media/project-compute-tier-direct.png)
57+
58+
59+
## Accessing Azure Notebooks Files from DSVM
60+
61+
To preserve parity of file paths with free compute you are only able to open one project at a time on a DSVM. If you would like to open a new project after opening a project, you will need to shutdown the open project first.
62+
63+
When a project is run on a VM, the files are mounted on the root directory of the Jupyter server (the directory shown in JupyterHub), replacing the default Azure Notebooks files. When you shut down the VM using the **Shutdown** button on the notebook UI, Azure Notebooks restores the default files.
64+
65+
![Shutdown button in Azure Notebooks](media/shutdown.png)

0 commit comments

Comments
 (0)