1

Topic: Replication and bases.

Good afternoon.
Read horror stories from . and cases from life there was a question: how competently to organize regular  replicable basis?
[quote =] "Bin-broad gull and replications - For replication" mysqldump "it is not intended, for this purpose there is a bin-broad gull (- log-bin)"

And on the contrary:
[quote =] "1. Recovery from  is possible only when they were conducted from the very beginning. Or if there is a mold of an initiating state of a database after which broad gulls started to be conducted.
2. On the loaded servers  quickly kill space on a disk, therefore they are stored not by everything, and a quantity of the last. Previous or are packed, or at all are deleted. It is clear that in the latter case to recover it will be possible nothing.
3. One more assignment  - replication. If there is a remark, unconditionally it is easier to recover the data from it, than from . "

And 1000 more and 1 judgements on everything, but not about the specific: how on replicable basis (on one master and ) it is correct to me to make dump.sql?!
It is necessary to do it on , or on the master? If on the master stopping replication, or "on live"?
Essence - I want to have near by "the fresh" archive ready at any moment to

SOURCE./dump.sql

How such task is competently organized? If someone has a link to appropriate article - will be grateful! Meanwhile only I plan to adjust and tear  with replication, I collect recipes and "good experts"!))