PackRecorder > Capturing Large Packages

Whilst the capturing and comparison modules are fast on their own, certain settings of the scanning should be fine-tuned when a big package (more than 50 000 files or so) is to be repackaged. This chapter summarizes how to identify the bottlenecks and improve the performance of scanning.

Permissions Scanning

By default, RayPack is scanning the system changes and looking for changes in the file/registry permissions. Scanning of permissions requires additional time and resources, which - depending on the machine, snapshot, and the package - may negatively impact on the performance and snapshot comparison.

If a package is known to contain no custom permissions, it is recommended to disable the unnecessary scanning options. For most of packages, the permissions are not changed by the installer and thus the option can be disabled without any loss of functionality.

Unnecessary Scanned Areas

If it is possible, the list of scanned resources (drives, registry hives) should be limited to absolute minimum. Drives that are not to be affected by the application (usually non-system drives) are increasing the time needed to finish a snapshot, and make them bigger, which results in more time required for the analysis.

For most of packages, a combination of C:\ drive and HKEY_LOCAL_MACHINE and HKEY_CURRENT_USER node is sufficient to perform fast and reliable scanning.

Ignoring Resources

Excluded resources are still scanned and copied to the package folder so that it is possible to revert the initial exclusions in the future without the need to repackage again. This feature has a negative impact on the performance. In order to skip analyzing and copying these resources, adjust the ignore list according to section Ignoring certain resources when repackaging.

Computer Parameters

Regardless of all settings and fine-tuning of the snapshotting, it is a general rule of thumb that the better the machine, the faster the scanning will be. Since the repackaging is usually done on a virtual machine, it is important to assign as much RAM memory and as many CPU cores as possible. Because scanning requires a lot of I/O operations, it is a good choice to store the virtual machine files on an SSD drive.