[NIFI-7002] Handle large files internally in EC toolkit

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.10.0
• Fix Version/s: 1.13.0
• Component/s: Tools and Build
• Labels: encryption heap security streaming toolkit
• Assignee: Unassigned
• Reporter: Andy LoPresto
• Created: 09/Jan/20 18:01
• Updated: 28/Sep/21 16:19
• Resolved: 28/Sep/21 16:19

Description

The EC toolkit cannot handle large flow definition files (>100 MB uncompressed) due to its use of String to hold the XML content internally through frequent manipulations. As Strings are immutable in Java, this causes a lot of heap usage and duplication. The read/write to the file system also operates on the String directly rather than perform a streaming read/write.

We should change the internal representation of the file content (for all files) to be a BufferedInputStream or other similar stream representation rather than a String. We should test with very large files (1 GB) and clearly provide instructions and warnings if the file size exceeds that.

The logic of value encryption/replacement will have to change to handle this approach.

Issue Links

• is fixed by: NIFI-6999 Encrypt Config Toolkit fails on very large flow.xml.gz files