#!/bin/tcsh

# First grab the bulk of the tables.  Skip the big ones.
(mysqldump --no-data -uroot -pxxx123 -hpablo mydb alerts alerts_daily candles_5m candles_d; mysqldump --opt --single-transaction --master-data -u root -pxxx123 -h wally-balls mydb `mysql -u root -pxxx123 -e "SHOW TABLES" -h wally-balls mydb | grep -v '^\(alerts\|candles\)' | grep -v Tables_in_mydb`)|gzip -1 > /tmp/main_mydb.sql.gz

# Now grab the smaller parts of the big data
mysqldump -uroot -pxxx123 -hpablo --opt --single-transaction --skip-add-locks mydb alerts_daily candles_5m candles_d | gzip -1 > /tmp/big_mydb.sql.gz

# Finally, split the alerts into different tasks.  Respect the same caching rules as the alerts server.
./backup_helper.tcl

# notice the following options to mysqldump
#
# --no-data - We always start with the shell of the big tables.  That is
#             espeically necessary for the alerts table because of the way we
#             save and restore that data.
# --opt - This seems to be a default now.  In the old days you would crash if
#         you tried to dump a big table without this.
# --single-transaction - This is *not* optional.  It is required for use with
#                        --master-data and it is required because the default
#                        type of locking would lock up the server for too long.
# --master-data - This is not present in old versions of mysql.  (Drama does
#                 not have this, but luckily it is a feature of the client
#                 not the server.)  This is the only reasonable way to copy
#                 data to a slave.
# --skip-add-locks - By default the program will lock the table before it 
#                    starts restoring the data, and unlocks it at the end. 
#                    Among other things, this is supposed to give you a little
#                    speed.  I avoid that for the big tables.  Without the
#                    locks, I can check the status of the table with statements
#                    like SELECT COUNT(*) or SELECT MAX(id).
# --no-create-info - I use this on the alerts table because we split it up into
#                    several pieces.  Otherwise each time you started a new
#                    piece, it would overwrite the old ones, rather than adding
#                    to what we already have.

# Note:  You need to start restoring main_mydb.sql.gz before you restore any of the
# other files.  The beginning of main_mydb.sql.gz includes the table definitions for
# some of the other files.  If you try to restore two files at once, each will run
# at approximately the same speed as if they were one at a time, doubling the speed of
# the entire operation.  If you try to restore more than two files at a time, you will
# quickly run into diminishing returns.