Hi, On 2025-04-10 13:50:36 +0000, Bertrand Drouvot wrote: > Thanks for the patch! That sounds like a great addition. I was doing some > tests and did not see any issues. Also while doing the tests I thouhgt that it > could be useful to evict only from a subset of NUMA nodes (now that NUMA > awareness is in). We'd need to figure out what to do for buffers that are > spread > across NUMA nodes though. > > Does that sound like an idea worth to spend time on? (If so, I'd be happy to > work > on it).
I'm not sure that's common enough to warrant its own function. You can do that with pg_buffercache_evict(), it'll be slower than pg_buffercache_evict_all(), but given that determining the numa node already is somewhat expensive, I'm not sure it's going to make that big a difference. Greetings, Andres Freund
