from reading around and poking in our own svn I found these tidbits from the google gods related to your svn errors
increase the locker entries in /path/to/repository/db/DB_CONFIG
convert to fsfs instead of bdb
$ svnadmin create –fs-type fsfs /path/to/new/repository
$ svnadmin dump /path/to/repository | svnadmin load /path/to/new/repository
<jamest>i guess we outgrew the berkley db settings for subversion
<jamest>so I had to bump the #s up for the logs
<jamest>In a different test repository when I had this problem I bumped up the
<jamest> DB_CONFIG values to look like this:
<jamest>set_cachesize 0 8388608 8
<jamest> set_lg_regionmax 524288
<jamest> set_lg_bsize 2097152
<jamest> set_lg_max 8388608