Re: fdatasync performance problem with large number of DB files

From: Justin Pryzby <pryzby(at)telsasoft(dot)com>
To: Thomas Munro <thomas(dot)munro(at)gmail(dot)com>
Cc: David Steele <david(at)pgmasters(dot)net>, Fujii Masao <masao(dot)fujii(at)oss(dot)nttdata(dot)com>, Bruce Momjian <bruce(at)momjian(dot)us>, Paul Guo <guopa(at)vmware(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Michael Brown <michael(dot)brown(at)discourse(dot)org>, pgsql-hackers(at)postgresql(dot)org, Michael Paquier <michael(at)paquier(dot)xyz>
Subject: Re: fdatasync performance problem with large number of DB files
Date: 2021-06-18 01:11:01
Message-ID: 20210618011101.GQ31772@telsasoft.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Thomas, could you comment on this ?

On Sat, May 29, 2021 at 02:23:21PM -0500, Justin Pryzby wrote:
> On Tue, May 25, 2021 at 07:13:59PM -0500, Justin Pryzby wrote:
> > On Sat, Mar 20, 2021 at 12:16:27PM +1300, Thomas Munro wrote:
> > > > > + {
> > > > > + {"recovery_init_sync_method", PGC_POSTMASTER, ERROR_HANDLING_OPTIONS,
> > > > > + gettext_noop("Sets the method for synchronizing the data directory before crash recovery."),
> > > > > + },
> >
> > Is there any reason why this can't be PGC_SIGHUP ?
> > (Same as restart_after_crash, remove_temp_files_after_crash)
>
> I can't see any reason why this is nontrivial.
> What about data_sync_retry?
>
> commit 2d2d2e10f99548c486b50a1ce095437d558e8649
> Author: Justin Pryzby <pryzbyj(at)telsasoft(dot)com>
> Date: Sat May 29 13:41:14 2021 -0500
>
> Change recovery_init_sync_method to PGC_SIGHUP..
>
> The setting has no effect except during startup.
> But it's nice to be able to change the setting dynamically, which is expected
> to be pretty useful to an admin following crash recovery when turning the
> service off and on again is not so appealing.
>
> See also: 2941138e6, 61752afb2
>
> diff --git a/doc/src/sgml/config.sgml b/doc/src/sgml/config.sgml
> index d8c0fd3315..ab9916eac5 100644
> --- a/doc/src/sgml/config.sgml
> +++ b/doc/src/sgml/config.sgml
> @@ -9950,7 +9950,8 @@ dynamic_library_path = 'C:\tools\postgresql;H:\my_project\lib;$libdir'
> appear only in kernel logs.
> </para>
> <para>
> - This parameter can only be set at server start.
> + This parameter can only be set in the <filename>postgresql.conf</filename>
> + file or on the server command line.
> </para>
> </listitem>
> </varlistentry>
> diff --git a/src/backend/utils/misc/guc.c b/src/backend/utils/misc/guc.c
> index 87bc688704..796b4e83ce 100644
> --- a/src/backend/utils/misc/guc.c
> +++ b/src/backend/utils/misc/guc.c
> @@ -4945,7 +4945,7 @@ static struct config_enum ConfigureNamesEnum[] =
> },
>
> {
> - {"recovery_init_sync_method", PGC_POSTMASTER, ERROR_HANDLING_OPTIONS,
> + {"recovery_init_sync_method", PGC_SIGHUP, ERROR_HANDLING_OPTIONS,
> gettext_noop("Sets the method for synchronizing the data directory before crash recovery."),
> },
> &recovery_init_sync_method,
> diff --git a/src/backend/utils/misc/postgresql.conf.sample b/src/backend/utils/misc/postgresql.conf.sample
> index ddbb6dc2be..9c4c4a9eec 100644
> --- a/src/backend/utils/misc/postgresql.conf.sample
> +++ b/src/backend/utils/misc/postgresql.conf.sample
> @@ -774,7 +774,6 @@
> # data?
> # (change requires restart)
> #recovery_init_sync_method = fsync # fsync, syncfs (Linux 5.8+)
> - # (change requires restart)
>
>
> #------------------------------------------------------------------------------

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Jeff Davis 2021-06-18 01:13:57 Re: Outdated replication protocol error?
Previous Message Andres Freund 2021-06-18 01:08:27 Re: Avoid call MaintainOldSnapshotTimeMapping, if old_snapshot_threshold is disabled.