Hi! I'm interested in the vacuum concurrently feature being inside the core, so will try to review patch set and give valuable feedback. For now, just a few little thoughts..
One more thing is about pg_squeeze background workers. They act in an autovacuum-like fashion, aren't they? Maybe we can support this kind of relation processing in core too?
I don't think it is necessary when this feature will be an internal feature.
I agree so this feature is very important, I proposed it (and I very happy so Tonda implemented it), but I am not sure, if usage of this should be automatized, and if it should be, then
a) probably autovacuum should do,
b) we can move a discussion after vacuum full concurrently will be merged to upstream, please. Isn't very practical to have too many open targets.