Simon Riggs wrote: > > > We *must* make CREATE INDEX CONCURRENTLY work with HOT. The good news is > I think we can without significant difficulty. >
Yeah, I think CREATE INDEX CONCURRENTLY is much easier to solve. Though I am not completely convinced that we can do that without much changes to CREATE INDEX CONCURRENTLY logic. For example, I believe we still need to lock out HOT-updates before we start CREATE INDEX CONCURRENTLY. Otherwise we might end up creating two paths to the same tuple in the new index. Say, we have a table with two columns (int a, int b). We have an index on 'a' and building another index on 'b'. We got a tuple (10, 20) in the heap. In the first phase of CREATE INDEX CONCURRENTLY, this tuple would be indexed. If the tuple is HOT-updated to (10, 30) before the first phase ends, the updated tuple would again get indexed in the second phase. This would lead to two paths to the latest visible tuple from the new index. Am I missing something in your design that stops this from happening ? Thanks, Pavan -- EnterpriseDB http://www.enterprisedb.com ---------------------------(end of broadcast)--------------------------- TIP 9: In versions below 8.0, the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match