A secondary node fails to apply some transaction. Instead of injecting empty transactions, and try to get it going, I would like to re-sync/re-do the whole secondary node.
What are my options? Shall I wipe the disk and let it re-join so that it copies over everything new? Is there a magical trick using SQL statements which would let me do that without touching disk files?
For reasons, the data dir is mounted, so I cannot simply move it away. I can move of course all files. (yes, move, I play it save..).
What are my options? Shall I wipe the disk and let it re-join so that it copies over everything new? Is there a magical trick using SQL statements which would let me do that without touching disk files?
For reasons, the data dir is mounted, so I cannot simply move it away. I can move of course all files. (yes, move, I play it save..).