On 10/13/18, Amit Kapila <amit.kapila16@gmail.com> wrote:
> On Sat, Oct 6, 2018 at 12:17 AM John Naylor <jcnaylor@gmail.com> wrote:
>> -For normal mode, I taught fsm_set_and_search() to switch to a
>> non-extending buffer call, but the biggest missing piece is WAL
>> replay.
>>
>
> fsm_set_and_search()
> {
> ..
> + /*
> + * For heaps we prevent extension of the FSM unless the number of pages
> + * exceeds
> HEAP_FSM_EXTENSION_THRESHOLD. For tables that don't already
> + * have a FSM, this will save an inode and a few kB
> of space.
> + * For sane threshold values, the FSM address will be zero, so we
> + * don't bother dealing with
> anything else.
> + */
> + if (rel->rd_rel->relkind == RELKIND_RELATION
> + && addr.logpageno == 0)
>
> I am not sure if this is a solid way to avoid creating FSM. What if
> fsm_set_and_search gets called for the level other than 0?
Thanks for taking a look. As for levels other than 0, I think that
only happens when fsm_set_and_search() is called by fsm_search(),
which will not cause extension.
> Also,
> when the relation has blocks more than HEAP_FSM_EXTENSION_THRESHOLD,
> then first time when vacuum will try to record the free space in the
> page, won't it skip recording free space for first
> HEAP_FSM_EXTENSION_THRESHOLD pages?
Hmm, that's a good point.
> I think you have found a good way to avoid creating FSM, but can't we
> use some simpler technique like if the FSM fork for a relation doesn't
> exist, then check the heapblk number for which we try to update the
> FSM and if it is lesser than HEAP_FSM_EXTENSION_THRESHOLD, then avoid
> creating the FSM.
I think I see what you mean, but to avoid the vacuum problem you just
mentioned, we'd need to check the relation size, too. I've attached an
unpolished revision to do this. It seems to work, but I haven't tested
the vacuum issue yet. I'll do that and some COPY performance testing
in the next day or so. There's a bit more repetition than I would
like, so I'm not sure it's simpler - perhaps RecordPageWithFreeSpace()
could be turned into a wrapper around RecordAndGetPageWithFreeSpace().
Also new in this version, some non-functional improvements to hio.c:
-debugging calls that are #ifdef'd out.
-move some code out into a function instead of adding another goto.
>> I couldn't find a non-extending equivalent of
>> XLogReadBufferExtended(), so I might have to create one.
>>
>
> I think it would be better if we can find a common way to avoid
> creating FSM both during DO and REDO time. It might be possible if
> somethin like what I have said above is feasible.
That would be ideal.
-John Naylor
