Gurobi Remote Services Release Notes v11.0.0
You have successfully installed version 11.0.0 of the Gurobi Remote Services.
Gurobi Remote Services is a set of Gurobi features that allows a cluster of one or more machines to perform Gurobi computations on behalf of other machines. The key components of Remote Services are:
- Compute Server, which allows you to offload all Gurobi computations from a client machine onto a remote cluster.
- Distributed Workers, which can be used to perform parallel computation on multiple machines.
- The Cluster Manager, an application server that provides secured access to your Remote Services cluster, as well as providing a Web User Interface and a command-line tool that make it easier to manage and monitor your cluster. The Cluster Manager uses a database to store user accounts, API keys, jobs, batches, and settings. You can use Compute Servers without a Cluster Manager – we refer to that as a self-managed cluster. However, we recommended you use a Cluster Manager to get the rich set of features that it provides.
Â
Obtaining Your License
To manage your Gurobi licenses, you will need to visit the Gurobi User Portal. If you are a commercial user under maintenance, you should see your Gurobi 11 licenses under the Licenses section. If you would like to request a free academic license or other licenses, you can do so from the Request License section. Once you have a license on the Gurobi User Portal, you will need to follow the instructions for installing or downloading a license by clicking on the Install
or Open
button next to your license.
Â
Supported Platforms for Remote Services
Platform | Operating System | Notes |
---|---|---|
Windows 64-bit (win64) | Windows 10, 11, Windows Server 2016, 2019, 2022 | Â |
Linux® x86-64 64-bit (linux64) | Red Hat® Enterprise Linux 8, 9 |  |
SUSE® Enterprise Linux 15 | ||
Ubuntu® 20.04, 22.04 | ||
Amazon Linux 2 | ||
Mac OS 64-bit (macos_universal2) | 12 (Monterey), 13 (Ventura), 14 (Sonoma) | Â |
Linux® arm64 64-bit (linux64) | Red Hat® Enterprise Linux 8, 9 |  |
SUSE® Enterprise Linux 15 | ||
Ubuntu® 20.04, 22.04 | ||
Amazon Linux 2 |
Â
Supported Browsers by the Cluster Manager Web UI
Browser | Minimum Version | Notes |
---|---|---|
Chrome | 87 | Â |
Firefox | 87 | Â |
Edge | 44 | Â |
Safari | 14 | Â |
Â
Supported Databases by the Cluster Manager
Database | Version | Notes |
---|---|---|
MongoDB | 4.4 and up | Â |
Amazon DocumentDB | 4.0 and 5.0 | Â |
Azure Cosmos DB | 4.2 | Â |
Â
New or Improved Features
Global Updates
- Gurobi Compute Server now supports Gurobi Optimizer 11.0.x, in addition to past releases: 10.0.x, and 9.5.x. Older versions are no longer supported and packaged in the compute server.
Cluster Manager
- The Cluster Manager can now be integrated with a SAML 2.0 directory. The system administrator can specify connection parameters and account mapping. Once activated, users will be given access based on the user accounts defined in the SAML directory. Accounts with a system administrator role can log in using their local passwords, which enables them to administer the cluster even if there is a problem with the SAML configuration. System administrators will need to log in using the special login page by accessing the
/login/admin
link at the top of the main login page. System accounts (i.e., accounts with no interactive login) can always gain access using API keys.When upgrading an existing Cluster Manager to SAML 2.0 authentication, the system administrator will need to check existing usernames. The username attribute mapping will be used to associate existing users. Any user not found in the system will be treated as a new user. If it is necessary for existing users to match the format of new usernames, system administrators can rename the users to the desired format in the user table.
SAML 2.0 authentication is also supported with the
grbcluster login
command line tool. By default,grbcluster
will automatically launch a web browser on the local machine and remain in a waiting state until the authentication process is successfully completed. Alternatively, when using the--no-browser
option,grbcluster
will display a unique URL and a Device ID Code. This URL and code can be used to carry out the authentication process on any machine that has access to the Cluster Manager. - The Cluster Manager now supports a new configuration parameter to control username case sensitivity. By default, case sensitivity is enabled so that users with different uppercase and lowercase letters are handled as different users. For example,
JohnDoe
andjohndoe
are different users by default. If case sensitivity is disabled, usernames will match even if some letters are uppercase or lowercase. For example,JohnDoe
andjohndoe
will be the same user. When using SAML 2.0 or LDAP as a user repository, you need to use the setting supported by the directory.When disabling the case-sensitivity configuration, you need to make sure that existing usernames do not generate conflicts. For example, if you had
JohnDoe
andjohndoe
as different users, then disabling case-sensitivity will lead to conflicts with these users. In this case, rename or merge the users accordingly before changing the configuration. - The look-and-feel of the Cluster Manager WebUI has been upgraded, and the display of dates has been improved. The user can switch between the local time zone and UTC to display the dates by accessing the user menu. The selected timezone is indicated in column headers or in formatted dates.
- Security of the Cluster Manager and Compute Server has been improved by supporting the latest security policy HTTP headers, including
Content-Security-Policy
,Referrer-Policy
andPermissions-Policy
. - A new configuration parameter enables the system administrator to tune the history cleanup batch size. When jobs have reached their expiration limit, they will be deleted by blocks. This parameter configures the block size for the deletion process. This needs to be tuned according to the usage pattern. If the block size is too small, the process may not be able to keep up with the deletion. If the block size is too large, it may overload the database.
Database
- The Cluster Manager now supports Microsoft Azure CosmosDB 4.2 for more flexibility when deploying your cluster to Microsoft Azure.
- It is now possible to check and recreate missing indexes in the database using the command
grb_rsm database --fix-indexes
. If database indexes were not properly created, it could lead to performance issues and could impact the overall behavior of the Cluster Manager. It is recommended to run this command after restoring a database from a backup or doing some maintenance of the database. - When installing the 11.0 Cluster Manager on an existing deployment, the database will be automatically migrated to schema version 1.1.2. During that process, new fields will be added and indexes will be dropped and recreated. No data will be deleted, and the migration is a safe process. However, we do recommend that you perform a backup before starting the migration for a critical production system.
Compute Server
- Several limits can now be enforced server side for all optimization jobs using the following parameters:
MEMLIMIT
: Limits the total amount of memory. If more is needed, Gurobi will fail with anOUT_OF_MEMORY
error. Note that it is not possible to retrieve solution information after this termination error.SOFTMEMLIMIT
: Limits the total amount of memory available to Gurobi. If more is needed, Gurobi will terminate with aMEM_LIMIT
status code, leading to a graceful exit of the optimization, such that it is possible to retrieve solution information afterwards or (in the case of a MIP solve) resume the optimization.TIMELIMIT
: Limits the total time expended during optimization. If the limit is reached, the optimization will return theTIME_LIMIT
status.
Â
License Agreement
Note that this software is covered by the Gurobi End User License Agreement. By completing the Gurobi installation process and using the software, you are accepting the terms of this agreement.
Thank you for using Gurobi products!
Â