Re: [HACKERS] [PATCH] Vacuum: Update FSM more frequently

Поиск
Список
Период
Сортировка
От Masahiko Sawada
Тема Re: [HACKERS] [PATCH] Vacuum: Update FSM more frequently
Дата
Msg-id CAD21AoC6MjJafgz3TC0jk+jaWF0h=ySd=omCj08=cUQU3Pv3WA@mail.gmail.com
обсуждение исходный текст
Ответ на Re: [HACKERS] [PATCH] Vacuum: Update FSM more frequently  (Claudio Freire <klaussfreire@gmail.com>)
Ответы Re: [HACKERS] [PATCH] Vacuum: Update FSM more frequently  (Masahiko Sawada <sawada.mshk@gmail.com>)
Список pgsql-hackers
On Fri, Mar 2, 2018 at 10:50 PM, Claudio Freire <klaussfreire@gmail.com> wrote:
> On Fri, Mar 2, 2018 at 10:47 AM, Claudio Freire <klaussfreire@gmail.com> wrote:
>> On Fri, Mar 2, 2018 at 7:38 AM, Masahiko Sawada <sawada.mshk@gmail.com> wrote:
>>> Thank you for updating the patches!
>>>
>>> +/*
>>> + * When a table has no indexes, vacuum the FSM at most every this
>>> + * many dirty pages. With a default page size of 8kb, this value
>>> + * basically means 8GB of dirtied pages.
>>> + */
>>> +#define VACUUM_FSM_EVERY_PAGES 1048576
>>>
>>> Is it better if we vacuum fsm every 8GB regardless of the block size?
>>> Because an installation that uses >8GB block size is likely to have
>>> the pages less than what an 8GB block size installation has, the
>>> current patch might lead to delay fsm vacuum. What do you think? If
>>> it's better, we can define it as follows.
>>>
>>> #define VACUUM_FSM_EVERY_PAGES ((8 * 1024 * 1024) / BLCKSZ)
>>
>> I honestly don't know. I've never tweaked page size, and know nobody
>> that did, so I have no experience with those installations.
>>
>> Lets go with your proposal.
>>
>>>
>>> --
>>> @@ -470,7 +484,9 @@ lazy_scan_heap(Relation onerel, int options,
>>> LVRelStats *vacrelstats,
>>>      TransactionId relfrozenxid = onerel->rd_rel->relfrozenxid;
>>>      TransactionId relminmxid = onerel->rd_rel->relminmxid;
>>>      BlockNumber empty_pages,
>>> -                vacuumed_pages;
>>> +                vacuumed_pages,
>>> +                fsm_updated_pages,
>>> +                vacuum_fsm_every_pages;
>>>      double        num_tuples,
>>>                  tups_vacuumed,
>>>                  nkeep,
>>>
>>> Regarding fsm_updated_pages variable name, I think it's better to be
>>> freespace_updated_pages because we actually counts the page updated
>>> its freespace, not fsm.
>>
>> Good point.
>>
>> New version attached.
>
> Sorry, forgot to actually squash. Now the real new version is attached.

Thank you for updating. I think the patches has enough discussion and
quality, so can go to the committer's review. I've marked them as
Ready for Committer.

Regards,

--
Masahiko Sawada
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center


В списке pgsql-hackers по дате отправления:

Предыдущее
От: David Rowley
Дата:
Сообщение: Re: [HACKERS] Removing LEFT JOINs in more cases
Следующее
От: Masahiko Sawada
Дата:
Сообщение: Re: [HACKERS] [PATCH] Vacuum: Update FSM more frequently