If memory serves, WAL is used for data streams if enabled. Replication is done at storage layer within region. Specific question can be sent to SA or Product team through your account representative.
WAL buffers are used to hold write ahead log (WAL) records that aren’t yet written to storage. The size of the WAL buffer cache is controlled by the wal_buffers setting. Aurora uses a log-based storage engine and changes are sent to storage nodes for persistence. Given the difference in how writes are handled by the Aurora storage engine, this parameter should be left unchanged when using Aurora PostgreSQL.
Sure Julien. I will ask Amazon too. My intention was just to gather this info IN CASE if someone has already an idea on this. Does that mean that I should not be posting Aurora postgres related questions/ doubts here? Please correct me if I am wrong!
Thanks,
Goti
On Sun, Feb 6, 2022 at 6:55 PM Julien Rouhaud <rjuju123@gmail.com> wrote:
Hi,
On Sun, Feb 06, 2022 at 05:24:37PM +0530, Goti wrote: > > Can someone shed some light as to whether aurora postgres uses wal buffers > or not? Given the way the writes are handled differently, I would like to > understand the same while it is being said in the documentation not to > change the parameter from the default value of 16MB.
It's unlikely that any one here can answer the question. You should ask Amazon, not us.
