Re: mysql 5.5 mysqldump

[Shiva V]

Rick,

Thanks for the update but i definitely have a problem with your theory. There is always some processing going on with the database and you can't take a full dump often (might work with small databases but what if the database is too large?). Most of the RDMBS systems now have incremental backups implementation now and they are the most effective way of backing up the systems without wasting tape resources and database resources too. So, incremental backups SHOULD BE DONE. And i don't think you will have any point in time recovery without any sort of incremental strategies implemented.

I know that you can get a consistent backup by stopping the slave and take a snapshot but i already told that the only tool i have at my disposal is "mysqldump" and stopping the slave for the backup doesn't make any sense to me as "mysql" provides "--single-transaction" option to me, which still gives me a consistent backup and i have to use the binary logs to make it point in time.

I am a Oracle DBA and just got into MySQL administration, so in the learning process i am going to compare (even it is not gonna work) but since MySQL is a RDBMS too, it should be same at the architecture wise (with not many feature, that's why it is cheap).

I was hoping to get an idea from an architectural point of view and never from a documentation/syntax perspective. I am going to test my theory and will let you guys know my results.

Thanks,
Shiva