catch size and performance

Hello:

I, running a openldap setup with one provider and 3 consumers , I am seeing intermittent problems of replication not happening "Until" the consumers are re started . This cures the problem. replication starts up and continues for a while .

The servers have very large amount of memory , 256 gigs ..

I configured the hdb backend 'DB_CONFIG' , to have 4 gigs of catch , could this be causing the replication problem ?

whats the optional size that keeps openldap happy ?

thanks

Greg