"Mikheev, Vadim" <vmikheev@SECTORBASE.COM> writes:
> Also, WAL doesn't prevent zero blocks in files after crash - I didn't
> want to fsync log on each new block allocation, - but this shouldn't
> be a problem (except of lost disk space), afair scans are smart to
> handle it, Tom?
This is OK for table files, unless someone's broken the code that will
auto-initialize a zero page when it comes across one.
I had a note to myself saying that zeroed pages in indexes may not be
OK, but I'm not sure if that's really a risk or not. btree, at least,
never does any linear scans of an index file; it can only visit pages
that are linked to by parent or sibling links. If we are careful that
we init a btree page before we modify its sibling(s) and parent, we
should be OK.
regards, tom lane
