| From: | "Bossart, Nathan" <bossartn(at)amazon(dot)com> |
|---|---|
| To: | "alvherre(at)alvh(dot)no-ip(dot)org" <alvherre(at)alvh(dot)no-ip(dot)org> |
| Cc: | Robert Haas <robertmhaas(at)gmail(dot)com>, Kyotaro Horiguchi <horikyota(dot)ntt(at)gmail(dot)com>, "x4mmm(at)yandex-team(dot)ru" <x4mmm(at)yandex-team(dot)ru>, "a(dot)lubennikova(at)postgrespro(dot)ru" <a(dot)lubennikova(at)postgrespro(dot)ru>, "hlinnaka(at)iki(dot)fi" <hlinnaka(at)iki(dot)fi>, "matsumura(dot)ryo(at)fujitsu(dot)com" <matsumura(dot)ryo(at)fujitsu(dot)com>, "masao(dot)fujii(at)gmail(dot)com" <masao(dot)fujii(at)gmail(dot)com>, "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: archive status ".ready" files may be created too early |
| Date: | 2021-08-23 16:48:24 |
| Message-ID: | 4268438D-343C-4831-8511-17B03CEA028E@amazon.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On 8/23/21, 9:33 AM, "alvherre(at)alvh(dot)no-ip(dot)org" <alvherre(at)alvh(dot)no-ip(dot)org> wrote:
> On 2021-Aug-23, Bossart, Nathan wrote:
>
>> Hm. My expectation would be that if there are no registrations, we
>> cannot create .ready files for the flushed segments. The scenario
>> where I can see that happening is when a record gets flushed to disk
>> prior to registration. In that case, we'll still eventually register
>> the record and wake up the WAL writer process, which will take care of
>> creating the .ready files that were missed earlier. Is there another
>> case you are thinking of where we could miss registration for a cross-
>> segment record altogether?
>
> I'm thinking of the case where no record cross segment boundaries ever.
Sorry, I'm still not following this one. If we skipped creating
.ready segments due to a crash, we rely on RemoveOldXlogFiles() to
create them as needed in the end-of-recovery checkpoint. If a record
fits perfectly in the end of a segment, we'll still register it as a
boundary for the next segment (hence why we use XLByteToSeg() instead
of XLByteToPrevSeg()). If database activity stops completely, there
shouldn't be anything to mark ready.
Nathan
| From | Date | Subject | |
|---|---|---|---|
| Next Message | alvherre@alvh.no-ip.org | 2021-08-23 17:14:09 | Re: archive status ".ready" files may be created too early |
| Previous Message | Pavel Stehule | 2021-08-23 16:44:24 | pretty slow merge-join due rescan? |