checkpoint_segments upgrade recommendation?

Started by Peter Eisentrautover 10 years ago5 messageshackers
Jump to latest
#1Peter Eisentraut
peter_e@gmx.net

The release notes say that checkpoint_segments has been replaced by
max_wal_size and min_wal_size, but there is no indication on how to
convert between the old and new settings. I think a lot of people will
have checkpoint_segments delicately tuned, so we should at least give
them a hint on how to carry that forward in spirit.

--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

#2Robert Haas
robertmhaas@gmail.com
In reply to: Peter Eisentraut (#1)
Re: checkpoint_segments upgrade recommendation?

On Fri, Oct 16, 2015 at 2:52 PM, Peter Eisentraut <peter_e@gmx.net> wrote:

The release notes say that checkpoint_segments has been replaced by
max_wal_size and min_wal_size, but there is no indication on how to
convert between the old and new settings. I think a lot of people will
have checkpoint_segments delicately tuned, so we should at least give
them a hint on how to carry that forward in spirit.

Yeah, it would be nice to have some guidance about that. But do we
know what the guidance should be?

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

#3Michael Paquier
michael@paquier.xyz
In reply to: Robert Haas (#2)
Re: checkpoint_segments upgrade recommendation?

On Sat, Oct 17, 2015 at 4:45 AM, Robert Haas <robertmhaas@gmail.com> wrote:

On Fri, Oct 16, 2015 at 2:52 PM, Peter Eisentraut <peter_e@gmx.net> wrote:

The release notes say that checkpoint_segments has been replaced by
max_wal_size and min_wal_size, but there is no indication on how to
convert between the old and new settings. I think a lot of people will
have checkpoint_segments delicately tuned, so we should at least give
them a hint on how to carry that forward in spirit.

Yeah, it would be nice to have some guidance about that. But do we
know what the guidance should be?

I think that we should just suggest a reverse formula of the maximum soft
limit of checkpoint_segments for max_wal_size in the release notes of 9.5,
basically:
(3 * your_old_checkpoint_segments + 1) * 16MB = max_wal_size
I am not sure it is worth mentioning that one needs to be be sure to keep
some extra room to handle potential spikes because that's not a hard limit,
but people who have already played with pg_xlog on a different partition
are already aware of that after tuning checkpoint_segments.
min_wal_size is a new parameter though, I don't think it matters much to
hint users for the transfer to the new configuration...
Regards,
--
Michael

#4Peter Eisentraut
peter_e@gmx.net
In reply to: Michael Paquier (#3)
Re: checkpoint_segments upgrade recommendation?

On 10/17/15 10:25 AM, Michael Paquier wrote:

I think that we should just suggest a reverse formula of the maximum
soft limit of checkpoint_segments for max_wal_size in the release notes
of 9.5, basically:
(3 * your_old_checkpoint_segments + 1) * 16MB = max_wal_size

How about this patch?

(Actually, I'd remove the + 1 to make the numbers come out rounder.)

Attachments:

0001-doc-Add-advice-on-updating-checkpoint_segments-to-ma.patchapplication/x-patch; name=0001-doc-Add-advice-on-updating-checkpoint_segments-to-ma.patchDownload+16-1
#5Michael Paquier
michael@paquier.xyz
In reply to: Peter Eisentraut (#4)
Re: checkpoint_segments upgrade recommendation?

On Wed, Oct 21, 2015 at 3:11 AM, Peter Eisentraut <peter_e@gmx.net> wrote:

On 10/17/15 10:25 AM, Michael Paquier wrote:

I think that we should just suggest a reverse formula of the maximum
soft limit of checkpoint_segments for max_wal_size in the release notes
of 9.5, basically:
(3 * your_old_checkpoint_segments + 1) * 16MB = max_wal_size

How about this patch?

(Actually, I'd remove the + 1 to make the numbers come out rounder.)

Removing the + 1 is fine for me.

+ been removed. Its place it taken by the new
"Its place is taken".

Other than those little things this looks fine to me.
Regards,
--
Michael