Hi, Thanks for the feedback!
> I think it would be good start point to use the same batch size of > CompactCheckpointerRequestQueue() and AbsorbSyncRequests() > So we’ll keep both batch processing and the request cap in place for now. > > > Right, but another point is to avoid lengthy holding of > > > CheckpointerCommLock. What do you think about that? > > > > I am not clear on this. Could you elaborate on it? > > See [1] for more detailed description of this. > Links. > 1. > https://www.postgresql.org/message-id/flat/db4534f83a22a29ab5ee2566ad86ca92%40postgrespro.ru I read the thread but didn’t find a specific explanation of avoiding long lock holds. My understanding is: when compaction processes a very large queue in one go, it holds CheckpointerCommLock the entire time, blocking all ForwardSyncRequest callers. Batch processing would release the lock after each chunk, allowing other backends to make progress. Is that correct?
