ckaczor/azuredatastudio
1
0
Fork 0
mirror of https://github.com/ckaczor/azuredatastudio.git synced 2026-02-16 10:58:30 -05:00
Code Issues Packages Projects Releases Wiki Activity

Initial AD support for BDCs (#6741)

Browse Source 
Partially working AD support for BDCs with some known issues
- Plumbed through kerberos support to Notebooks.
  - Using "gateway-0" for service temporarily as service endpoints API doesn't yet return correct DNS name. Will update in separate PR once available
- Plumbed kerberos auth to HDFS, Spark. Only partially working as we use same token on each call 
  - Will fix in separate PR, as this requires a refactor of WebHDFS library. Will need to either get new token every time or set a cookie, both of which require refactors
- Fixed error when Data Service node expansion failed and blocked all OE expansion
- Support for SqlToolsService change to use new cluster endpoints DMV
  -  Updated API to add new endpoints field to replace IP + port
  - Added logic to handle case where endpoints for Yarn, Grafana etc. are in the list
  - Sort list and use expected new localized strings

- Updated SqlToolsService to include support for new DMV
- Add "gateway-0" handling in Jupyter session as workaround for lack of domain names in endpoints list
This commit is contained in:
Kevin Cunnane
2019-08-14 18:09:41 -07:00
committed by GitHub
parent 4e8c06f36d
commit 52f8984a99
31 changed files with 639 additions and 189 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
extensions/notebook/src/common/utils.ts
View File

@@ -141,4 +141,4 @@ export function isEditorTitleFree(title: string): boolean {
let hasTextDoc = vscode.workspace.textDocuments.findIndex(doc => doc.isUntitled && doc.fileName === title) > -1;
let hasNotebookDoc = azdata.nb.notebookDocuments.findIndex(doc => doc.isUntitled && doc.fileName === title) > -1;
return !hasTextDoc && !hasNotebookDoc;
}
}

28
extensions/notebook/src/jupyter/jupyterSessionManager.ts
View File

@@ -55,10 +55,11 @@ const configBase = {
const KNOX_ENDPOINT_SERVER = 'host';
const KNOX_ENDPOINT_PORT = 'knoxport';
const KNOX_ENDPOINT_KNOX = 'knox';
const KNOX_ENDPOINT_GATEWAY = 'gateway';
const SQL_PROVIDER = 'MSSQL';
const USER = 'user';
const AUTHTYPE = 'authenticationType';
const INTEGRATED_AUTH = 'integrated';
const DEFAULT_CLUSTER_USER_NAME = 'root';
export class JupyterSessionManager implements nb.SessionManager {
@@ -242,13 +243,19 @@ export class JupyterSession implements nb.ISession {
//Update server info with bigdata endpoint - Unified Connection
if (connection.providerName === SQL_PROVIDER) {
let clusterEndpoint: utils.IEndpoint =
await this.getClusterEndpoint(connection.id, KNOX_ENDPOINT_KNOX) ||
await this.getClusterEndpoint(connection.id, KNOX_ENDPOINT_GATEWAY);
let clusterEndpoint: utils.IEndpoint = await this.getClusterEndpoint(connection.id, KNOX_ENDPOINT_GATEWAY);
if (!clusterEndpoint) {
return Promise.reject(new Error(localize('connectionNotValid', "Spark kernels require a connection to a SQL Server big data cluster master instance.")));
}
connection.options[KNOX_ENDPOINT_SERVER] = clusterEndpoint.ipAddress;
if (this.isIntegratedAuth(connection)) {
// Hack: for now, we need to use gateway-0 for integrated auth
let sqlDnsName: string = connection.options['server'].split(',')[0];
let parts = sqlDnsName.split('.');
parts[0] = 'gateway-0';
connection.options[KNOX_ENDPOINT_SERVER] = parts.join('.');
} else {
connection.options[KNOX_ENDPOINT_SERVER] = clusterEndpoint.ipAddress;
}
connection.options[KNOX_ENDPOINT_PORT] = clusterEndpoint.port;
connection.options[USER] = DEFAULT_CLUSTER_USER_NAME;
}
@@ -259,8 +266,9 @@ export class JupyterSession implements nb.ISession {
this.setHostAndPort(',', connection);
let server = Uri.parse(utils.getLivyUrl(connection.options[KNOX_ENDPOINT_SERVER], connection.options[KNOX_ENDPOINT_PORT])).toString();
let doNotCallChangeEndpointParams =
`%_do_not_call_change_endpoint --username=${connection.options[USER]} --password=${connection.options['password']} --server=${server} --auth=Basic_Access`;
let doNotCallChangeEndpointParams = this.isIntegratedAuth(connection) ?
`%_do_not_call_change_endpoint --server=${server} --auth=Kerberos`
: `%_do_not_call_change_endpoint --username=${connection.options[USER]} --password=${connection.options['password']} --server=${server} --auth=Basic_Access`;
let future = this.sessionImpl.kernel.requestExecute({
code: doNotCallChangeEndpointParams
}, true);
@@ -268,6 +276,10 @@ export class JupyterSession implements nb.ISession {
}
}
private isIntegratedAuth(connection: IConnectionProfile): boolean {
return connection.options[AUTHTYPE] && connection.options[AUTHTYPE].toLowerCase() === INTEGRATED_AUTH.toLowerCase();
}
private isSparkKernel(kernelName: string): boolean {
return kernelName && kernelName.toLowerCase().indexOf('spark') > -1;
}
@@ -330,4 +342,4 @@ interface ISparkMagicConfig {
}
}
};
}
}