Many thanks @Joy Jin. I have improved the solution by decreasing the interval from 10 seconds to 1 second to prevent cleaning backups better:
while test 0=0
if tmutil status | grep -q "BackupPhase = ThinningPostBackup"
say "Prevented cleaning Time Machine backups&...
None of the existing answers worked in my situation:
client: osx/macos 10.x (panther -> catalina)
server: debian 6 (squeeze), samba 3.5.6
Turns out the share in question didn't have the correct hosts allow IP range in place to match specific host (i.e. error message is very misleading).