Saving big (100GB+) directories with lots of files under subversion will most probably fail (due to timeout/exeeded RAM consumption & co). Instead of One Big Failing Transaction, commit your data in chunks.
#!/bin/bash
# First, commit directory structure
# find -type d | xargs -d '\n' svn add --depth=empty
# Then, commit all files in chunks
for i in {1..200}
do
echo "Welcome $i times"
svn status | grep "^\?" | awk '{print $2}' | head -n 1000 | xargs -d '\n' svn add
svn commit -m "With chunk $i/200"
done