LOBs/Replication

[Brett Wooldridge <br...@gmail.com>]

I have a question about replication as it pertains to LOBs (BLOBs and CLOBs).
According to the documentation...

If the master looses connection with the slave, "transactions are
allowed to continue
processing while the master tries to reconnect with the slave. Log
records generated
while the connection is down are buffered in main memory. If the log
buffer reaches
its size limit before the connection can be reestablished, the master
replication
functionality is stopped."

And the documentation for derby.replication.logBufferSize says the
maximum size of
the buffer is 1048576 (1MB).

This seems to imply that if I have a database in which I store LOBs
which are, for
example, 256K in size, and the connection between master and slave is severed,
I can perform 4 inserts or less before the master gives up.  Is this
so?  If this is in fact
the case, I would like to file a request that this limit be raised
considerably or eliminated
altogether.

I have two servers (master and slave) running 64-bit JVMs, 64GB of memory each,
SSD drives, connected by 10GbE fiber.  I would like to dedicate as
much memory as I
want to deal with a disconnect/resume scenario (to avoid the onerous
failover).  At an
insertion rate of 16 rows per second (~4MB), currently the setup would
tolerate a
connection interruption of a fraction of a second.  A 1GB buffer would
afford a connection
interruption of ~250 seconds (for example, rebooting the fiber switch).

Last question, aimed at devs, why does Derby even bother to buffer
logs in memory?
Can't it just keep an offset/marker into the transaction log files,
and replay transactions
from there, rather than buffering them in memory?

Regards,
Brett