Release notes
Updates for version 1.14.0
On January 17, 2023, we released Migrate Fit Assessment 1.14.0.
Features
Report applications fit to Google containers platforms
Added a new report category to help you focus assessment report on applications fit for Google containers platforms—Anthos, GKE, Autopilot and Cloud Run. When generating a report you can ask to include only assessment to containers platforms using the following command:
./mfit report --format html --target-platform containers --full >
REPORT_NAME.html
Fit assessment of JBoss and Apache 2 applications
The Migrate Fit Assessment tool can assess JBoss and Apache 2 applications fit for containerization to GKE, Anthos, and Cloud Run platforms. To learn more about JBoss and Apache 2 containerization, see the Migrate to Containers 1.14.0 release notes.
Guest level collection without the need for root permission
Updated the Migrate Fit Assessment collection scripts for Linux to enable
running them through sudo
.
The added capability allows you to run Linux guest
collection script with sudo
and without the need for a root user.
The Linux guest collection script will attempt to run using passwordless sudo
.
This change takes effect in all Linux guest collection methods, such as:
mfit discover ssh
mfit discover vsphere guest
mfit discover vsphere guest all
- When running the collection script manually and using
mfit discover import
.
If a passwordless sudoer credentials are supplied, the Linux guest collection
script uses sudo
, resulting in fuller, and more reliable collection results.
Note that sudoer permissions that require a password prompt are not supported for this feature. In this case, the script runs with the permissions of the original user.
New detailed assessment report added under experimental flag
A new design of the detailed assessment report is now available running the
command ./mfit report --format html --full > REPORT_NAME.html --experimental
.
The new design introduces enhanced filtering and text-based search to help you locate candidate workloads for modernization.
Changes
Deprecation of --user
flag in Linux collection script
Previously, running with a non-root user required supplying a --user
flag to
the Linux guest collection script. This is no longer required, and the --user
flag is ignored and entered a deprecation period.
Fixed
Renamed the discover ssh connect-timeout
flag
Renamed the flag called --timeout
in the discover ssh
command to
--connect-timeout
. The flag can be used with the native SSH client.
The new name removes the overlap with the global --timeout
flag.
Failure when providing embedded SSH client with no authentication
Fixed failure in case the discover ssh
command without authentication
parameters. When using the embedded SSH client and providing no identity file
and no password or an empty one, panic occurred instead of stating unsuccessful
authentication.
Error when failing to create a tar file in Linux guest collection
If the collection script fails to create the required tar file, it throws an error. Previously, a failure to create a tar in the Linux guest collection script was ignored. From this version, an error message is displayed, and the script exits with status code 1.
Saving guest collection tar file to different location
To provide a different location for saving the collection tar file, you can
use the --output
flag to save it to the specified path. Previously, arguments
supplied after the --output
flag were ignored due to parsing issues.
Updates for version 1.13.1
On November 3, 2022 we released Migrate Fit Assessment 1.13.1.
Changes
Deprecated JSON v0 reports
We introduced changes in the structure of the generated JSON file to expand
the scope of the assessment details report.
Starting with release 1.13.1, the default JSON output will use the file
structure version v1.
To generate a JSON file in the previous version, apply the flag --json-v0
.
./mfit report --format json --json-v0 <file.json>
The v0 JSON report is now deprecated. In Q1, 2023, v0 JSON report format will reach end of life. New releases of mFit will not support v0 JSON report generation.
Fixed
Decreased database file size
The SQLite database used by mFit has been optimized to reduce the disk space it consumes.
Fixed too many SQL parameters
error
Fixed an error that might occur when running discover ls
or report
commands
on an mFit installation with a very large number of discovered assets.
Updates for version 1.13.0
On October 18, 2022 we released Migrate Fit Assessment 1.13.0.
Changes
Enhanced vSphere guest collection at scale with scope option
You can now perform guest collection for a subset of VMs by applying a scope via
the --path
flag.
mfit discover vsphere guest all --url https://VSPHERE_URL --path dc-1/host/esx-1
Enhanced error reporting in RVTools import
RVTools discovery now identifies unsupported file formats and provides you with an error message on unsupported versions.
Fixed
Fixed RAM and storage size conversions in reports
Adjusted unit measures in reports to GiB instead of GB, inline with the internal units of measure.
Updates for version 1.12.1
Features
Perform guest collection at scale
You can now perform guest collections in parallel for all VMs managed by a single VMware vCenter. This comes in handy for collecting information from large quantities of VMs having the same credentials.
To collect guest-level information from multiple VMs, run:
mfit discover vsphere guest all --url https://VSPHERE_URL
Changes
Update to Cloud Run fit assessment
Cloud Run fit assessment has been updated to accommodate improvements done in Cloud Run resources limitations. Previous values were 4 CPUs, 16 GB of memory, and the new limits are 8 CPUs, 32 GB of memory. See details at Cloud Run Quotas and Limits.
Unify report and assess commands
We simplified the process of assessing data and generating a report, by
combining the two commands into one. In previous releases, you needed to first
perform an assessment ($ mfit assess
), and then you could generate a report on
the collected data. Now you can run the report command, and the assessment is
done automatically. The report
command now has an optional --target-platform
flag.
The available options are: gcp
, anthos
, anthosbm
and all
. Each option
specifies a different set of journeys to evaluate. If you don't specify a target
platform, all journeys are assessed.
gcp
assesses shift journeys to Compute Engine, Google Cloud VMware Engine, and containerization journeys to Google Kubernetes Engine, Google Kubernetes Engine Autopilot and Cloud Run.anthos
assesses shift to Anthos on bare metal and containerization journeys to Google Kubernetes Engine and Cloud Run.anthosbm
is equivalent toanthos
.all
assesses all possible journeys.
With this change, the report sample
command replaces the now deprecated assess sample
command.
Fixed
Fix mentions of private / public previews in html reports
In previous releases, any containerize target is shown as private preview. Now only Tomcat targets are shown as public preview.
mfit collection scripts cleanup temporary directory
In previous releases, collection scripts would create a temp directory on the guest, but didn't remove it. In the latest release, collection scripts automatically remove the temp directory.
Fix wrong RAM size conversions
We corrected a conversion where base 2 was mixed base 10 (MiB vs MB) and resulted in inaccurate conversions.
Windows guest collection
We fall back to non-WMI based collection methods for some data to avoid sporadic failures.
Updates for version 1.12.0
Changes
- Assessment of VMs to Anthos on bare metal Attach mode was removed from the executive report and detailed assessment report.
- Modified the default assessment of Anthos on bare metal as part of
the executive report. In order to assess VMs for Anthos on bare metal
you must run the command:
mfit assess --target-platform anthosbm
- You'll see a warning if you issue a CLI command that attempts to assess an empty DB.
- Updated the default behavior of
assess –target-platform
flag to use theall
option.
Fixed
- Fixed a bug in the executive summary section which caused inaccurate counts of high CPU and high memory VMs.
- Improved the sorting on the fit assessment full report, so that results sorted by rule ID are correctly sorted.
- Improved the fit assessment full report middle chart sorting so rows sorted from highest to lowest aggregated fit score.
- Fixed an incorrect regex in the Linux collect script affecting Tomcat assessment.
- Fixed a bug that prevented correct detection of centos OS version name and caused empty OS names on fit assessment full report.
- Fixed a bug which caused duplicated VM information when VMs were discovered multiple times.
- Improved handling of partially collected VM information which may lead to inconsistent data. Warn users when collecting an asset with the inconsistent VM information and skip storing the VM data.
Updates for version 1.11.3
Features
Fit assessment cost analysis - Use this new section on the fit assessment report to estimate the cost of running modernized workloads in Google Cloud. The report groups your VMs by modernization strategy and shows the estimated costs by strategy, as well as an estimated total cost. Cost analysis will help you decide whether a low-effort or high modernization strategy suits your needs. To view the cost benefits of a modernization journey, you need to run the collection scripts on an assessed VM. For more detailed instructions, see Discover and collect data.
Improved recommendation listing for fit journeys - The full fit assessment report now sorts journeys based on their fit and the sum severity of all failed conditions. The first journey in the list has the highest recommendation, with journeys of higher failure severity ranking lower.
Fixed
- Fixed an issue where features were only calculated when running the
discover
command.
Updates for version 1.11.2
Features
Report filter – A new filter was added on the fit assessment detailed report to enable filtering VMs which have fit for specific journeys. Once the filter is applied, only VMs which have a fit for the selected journey are listed.
StratoZone – Added mFIT collection script to StratoZone StratoProbe collector 5.0.2.1. The information required to make an assessment is collected by StratoProbe collector, and is made available for assessment for users that have deployed StratoProbe.
Fixed
Removed the fit score column from full report table for better readability.
Updated fit assessment summary report to resolve mismatch of percentage presented on the summary part and guidance part of the report.
Changed the severity information to use text instead of numeric value on the summary report.
Fixed wrong identification of OS which was marked as Other OS on fit assessment reports.
Changes
Added the option to purge the fit assessment local database by using the command
mfit discover purge-db
Added information on the existence of VMware tools in the CSV report. To allow users to automate guest level data collection through automated scripts.
Added a (
--minimal
) flag to Windows collection script which helps to minimize the amount of collected information and provide only the specific information required for assessing containerization to Windows IIS applications.Added the option to manually check for newer versions of the fit assessment tool by using the command
mfit version --check
.
Updates for version 1.11.1
- Bug fixes.
Updates for version 1.11.0
- Executive summary report – This release includes a new type of report, and a change in the default behavior. The new executive summary report gives you an overview of your modernization journeys. At a high level, the report sorts your modernization journeys into two different buckets: low effort versus high modernization. Looking at these two extremes, and the benefits and effort required, should help you make key business decisions. In addition, there are subsections of the report that show the fit and technical limitations of migrating VMs (lift and shift analysis), refactoring to containers, and on-premises VM modernization (via shifting or attaching VMs to Anthos clusters on bare metal). The default report type is the new executive summary report; in previous releases, the default report was a detailed report.
- Assess workloads for migrating to Google Cloud VMware Engine – The fit assessment tool added new assessment capability to assess workloads for lift and shift to Google Cloud VMware Engine.
Updates for version 1.10.0
The fit assessment tool for version 1.10.0 adds new functionality, including:
- Support for Autopilot, Cloud Run, and Compute Engine assessment
- Tomcat rules added for fit assessment
- Ability to acquire vSphere data from RVTools export
- Automatic version check
- Fit assessment tool configuration