Resizing chunk files in Syncrify
File comparison in Syncrify is done in 3 stages:
- Matching size and date of the file on both ends
- Matching MD5 checksum of the file
- Matching blocks
Block matching (Stage 3) occurs if a match is not found using the first two methods. This is the most time consuming stage. To reduce this time, Syncrify splits a large file into smaller chunks of 1TB *
each. For example, the original file is 10TB, Syncrify will try to split it up in 10 files of 1TB each. Once these files are split, MD5 checksum is computed to determine the chunks that match, hence avoiding block matching on the entire file.
The chunk size is set to 1TB*
by default and is configurable. Following steps demonstrate how to configure this value:
Prior to v3.1 the default chunk size was 8GB
- Locate SyncrifyClient.lax file in C:\Program Files\SyncrifyClient folder.
- Open this file in any editor
- Search for lax.nl.java.option.additional . If this line is not found, add a new line that reads:
- If the line already existing, append following towards the end.
- This sets the chunk size to 100GB. (107374182400 is the number of bytes in 100GB)
- Save the file
Specify a -D parameter as a command line argument.
java -Dmax.file.part=107374182400 -jar SyncrifyClient.jar
This will set the chunk size to 100GB