Multimaster replication with BLOBS

I am setting up multimaster (2 sites to start with) replication with 8.1.7 (both machines), running on a Sun Solaris 8 box and a x86 Solaris 8 box. Replication works fine with standard non-blob tables. However, when I try to replicate a table with BLOBs I get thousands of errors (something about blob fragmentation).

Is there something special I need to do to support blob replication? All tables have unique PK's, and values in init.ora for: shared_pool_size, compatible (8.1.7), global_names, open_links, distributed_transactions, job_queue_internval, job_queue_processes have been set to recommended or higher values. There is plenty of table space, etc. for all tables, indexes, and blobs.

Is there any documentation on BLOB replication (from someone other than Oracle)? I need a practical insight into what can / is going wrong here. Weird thought, but is there something that prevents replication of blobs between mixed endian machines?

HELP... (loosing hair by the moment)

Bryan Received on Fri Sep 07 2001 - 09:26:06 CDT