On Mon, Jan 25, 2021 at 2:11 PM Heikki Linnakangas <hlinn...@iki.fi> wrote: > > [patches 0001 - 0003] > > Makes sense.
Great. I committed the first one and will proceed with those as well. > > So, the CLOG page gets created when nextXid advances from the first > > value on the page to the second value on the page, not when it > > advances from the last value on the previous page to the first value > > on the current page. > Yes. It took me a moment to understand that explanation, though. I'd > phrase it something like "nextXid is the next XID that will be used, but > we want to set latest_page_number according to the last XID that's > already been used. So retreat by one." OK, updated the patch to use that language for the comment. > Having a separate FullTransactionIdToLatestPageNumber() function for > this seems like overkill to me. I initially thought so too, but it turned out to be pretty useful for writing debugging cross-checks and things, so I'm inclined to keep it even though I'm not at present proposing to commit any such debugging cross-checks. For example I tried making the main redo loop check whether XactCtl->shared->latest_page_number == FullTransactionIdToLatestPageNumber(nextXid) which turned out to be super-helpful in understanding things. -- Robert Haas EDB: http://www.enterprisedb.com
v2-0001-In-clog_redo-don-t-set-XactCtl-shared-latest_page.patch
Description: Binary data
v2-0002-In-TrimCLOG-don-t-reset-XactCtl-shared-latest_pag.patch
Description: Binary data
v2-0003-In-StartupCLOG-correct-an-off-by-one-error.patch
Description: Binary data
