Hi, > > This patch fails to consider that len may be bigger than MaxHeapTupleSize * > 0.98, which in this case triggers a reproducable > PANIC:
Good catch! I've adapted the patch with your suggested fix. > > One different question I have, though, is why we can't "just" teach vacuum > to clean up trailing unused line pointers. As in, can't we trim the line > pointer > array when vacuum detects that the trailing line pointers on the page are all > unused? > > The only documentation that I could find that this doesn't happen is in the > comment on PageIndexTupleDelete and PageRepairFragmentation, both not > very descriptive on why we can't shrink the page->pd_linp array. One is > "Unlike heap pages, we compact out the line pointer for the removed tuple." > (Jan. 2002), and the other is "It doesn't remove unused line pointers! Please > don't change this." (Oct. 2000), but I can't seem to find the documentation / > conversations on the implications that such shrinking would have. > This is an interesting alternative indeed. I also can't find any documentation/conversation about this and the message is rather cryptic. Hopefully someone on the list still remembers the reasoning behind this rather cryptic comment in PageRepairFragmentation. -Floris
v3-Allow-inserting-tuples-into-almost-empty-pages.patch
Description: v3-Allow-inserting-tuples-into-almost-empty-pages.patch