"distributed checkpoint"

Alvaro Herrera <alvherre@alvh.no-ip.org> writes:

Am I the only one who finds the phrase "distributed checkpointing" a bit
awkward? Would it be better if we used "time-distributed checkpointing"
instead?

Yeah, "distributed" has a bunch of connotations that are wrong for this
purpose.

I spent a bit of time with a thesaurus but didn't come up with anything
that seemed le mot juste. Best I could do was "spread checkpoint"
or "time-extended checkpoint". Anybody have a better idea?

regards, tom lane

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu, 06 Dec 2007 20:44:49 -0500
Tom Lane <tgl@sss.pgh.pa.us> wrote:

Alvaro Herrera <alvherre@alvh.no-ip.org> writes:

Am I the only one who finds the phrase "distributed checkpointing"
a bit awkward? Would it be better if we used "time-distributed
checkpointing" instead?

Yeah, "distributed" has a bunch of connotations that are wrong for
this purpose.

I spent a bit of time with a thesaurus but didn't come up with
anything that seemed le mot juste. Best I could do was "spread
checkpoint" or "time-extended checkpoint". Anybody have a better
idea?

balanced
gradual
extended (I see you mention time-extended but wouldn't time be implicit
based on the actual docs and thus we only need extended?)

Sincerely,

Joshua D. Drake

regards, tom lane

---------------------------(end of
broadcast)--------------------------- TIP 5: don't forget to increase
your free space map settings

- --

=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 24x7/Emergency: +1.800.492.2240
PostgreSQL solutions since 1997 http://www.commandprompt.com/
UNIQUE NOT NULL
Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFHWLvWATb/zqfZUUQRAkYAAJoCMHhtiWA6qxDG7U7UlfWsGBf+3QCeLwab
lzoTVeD8YvEme5M2TFi8NF8=
=c7bI
-----END PGP SIGNATURE-----

On Thu, Dec 06, 2007 at 07:19:44PM -0800, Joshua D. Drake wrote:

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu, 06 Dec 2007 20:44:49 -0500
Tom Lane <tgl@sss.pgh.pa.us> wrote:

Alvaro Herrera <alvherre@alvh.no-ip.org> writes:

Am I the only one who finds the phrase "distributed checkpointing"
a bit awkward? Would it be better if we used "time-distributed
checkpointing" instead?

Yeah, "distributed" has a bunch of connotations that are wrong for
this purpose.

I spent a bit of time with a thesaurus but didn't come up with
anything that seemed le mot juste. Best I could do was "spread
checkpoint" or "time-extended checkpoint". Anybody have a better
idea?

balanced
gradual
extended (I see you mention time-extended but wouldn't time be implicit
based on the actual docs and thus we only need extended?)

How about "smoothed?"

Cheers,
David.
--
David Fetter <david@fetter.org> http://fetter.org/
Phone: +1 415 235 3778 AIM: dfetter666 Yahoo!: dfetter
Skype: davidfetter XMPP: david.fetter@gmail.com

Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu, 6 Dec 2007 19:43:29 -0800
David Fetter <david@fetter.org> wrote:

On Thu, Dec 06, 2007 at 07:19:44PM -0800, Joshua D. Drake wrote:

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu, 06 Dec 2007 20:44:49 -0500
Tom Lane <tgl@sss.pgh.pa.us> wrote:

Alvaro Herrera <alvherre@alvh.no-ip.org> writes:

Am I the only one who finds the phrase "distributed
checkpointing" a bit awkward? Would it be better if we used
"time-distributed checkpointing" instead?

Yeah, "distributed" has a bunch of connotations that are wrong for
this purpose.

I spent a bit of time with a thesaurus but didn't come up with
anything that seemed le mot juste. Best I could do was "spread
checkpoint" or "time-extended checkpoint". Anybody have a better
idea?

balanced
gradual
extended (I see you mention time-extended but wouldn't time be
implicit based on the actual docs and thus we only need extended?)

How about "smoothed?"

Pre-emptive apologies:

silky
satin
butter

Man... that would rock silky_checkpoint = true ;)

:P

Joshua D. Drake

Cheers,
David.

- --

=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 24x7/Emergency: +1.800.492.2240
PostgreSQL solutions since 1997 http://www.commandprompt.com/
UNIQUE NOT NULL
Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFHWMNSATb/zqfZUUQRAt/yAKCEI2b7LTm7jM32Qx2HAM5NCZCh4QCfWAg5
qswiDmCrSWKH2RHCqPWw5j4=
=xmGU
-----END PGP SIGNATURE-----

David Fetter wrote:

balanced
gradual
extended (I see you mention time-extended but wouldn't time be implicit
based on the actual docs and thus we only need extended?)

How about "smoothed?"

perhaps we should call it Jacob checkpointing, then.

cheers

andrew

A Dijous 06 Desembre 2007, Alvaro Herrera va escriure:

Am I the only one who finds the phrase "distributed checkpointing" a bit
awkward? Would it be better if we used "time-distributed checkpointing"
instead?

The phrase is used in the release notes, but it's not used anywhere in
the main docs.

To me, non-native english speaker, time-distributed seems the clearer one.

--
Albert Cervera i Areny
http://www.NaN-tic.com

On Thu, 2007-12-06 at 19:43 -0800, David Fetter wrote:

On Thu, Dec 06, 2007 at 07:19:44PM -0800, Joshua D. Drake wrote:

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu, 06 Dec 2007 20:44:49 -0500
Tom Lane <tgl@sss.pgh.pa.us> wrote:

Alvaro Herrera <alvherre@alvh.no-ip.org> writes:

Am I the only one who finds the phrase "distributed checkpointing"
a bit awkward? Would it be better if we used "time-distributed
checkpointing" instead?

Yeah, "distributed" has a bunch of connotations that are wrong for
this purpose.

I spent a bit of time with a thesaurus but didn't come up with
anything that seemed le mot juste. Best I could do was "spread
checkpoint" or "time-extended checkpoint". Anybody have a better
idea?

balanced
gradual
extended (I see you mention time-extended but wouldn't time be implicit
based on the actual docs and thus we only need extended?)

How about "smoothed?"

Agreed

"Smoothed" makes a lot of sense for me. We used to have a checkpoint
spike, now we don't.

Perhaps we should say something like "time extended checkpoints provide
smoother (transaction?) response times"

--
Simon Riggs
2ndQuadrant http://www.2ndQuadrant.com

Ühel kenal päeval, R, 2007-12-07 kell 18:22, kirjutas Simon Riggs:

On Thu, 2007-12-06 at 19:43 -0800, David Fetter wrote:

On Thu, Dec 06, 2007 at 07:19:44PM -0800, Joshua D. Drake wrote:

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu, 06 Dec 2007 20:44:49 -0500
Tom Lane <tgl@sss.pgh.pa.us> wrote:

Alvaro Herrera <alvherre@alvh.no-ip.org> writes:

Am I the only one who finds the phrase "distributed checkpointing"
a bit awkward? Would it be better if we used "time-distributed
checkpointing" instead?

Yeah, "distributed" has a bunch of connotations that are wrong for
this purpose.

I spent a bit of time with a thesaurus but didn't come up with
anything that seemed le mot juste. Best I could do was "spread
checkpoint" or "time-extended checkpoint". Anybody have a better
idea?

balanced
gradual
extended (I see you mention time-extended but wouldn't time be implicit
based on the actual docs and thus we only need extended?)

How about "smoothed?"

Agreed

"Smoothed" makes a lot of sense for me. We used to have a checkpoint
spike, now we don't.

wide checkpoints ?

provide wide and low spikes :)

or even background checpoints ?

On Fri, 7 Dec 2007, Simon Riggs wrote:

"Smoothed" makes a lot of sense for me. We used to have a checkpoint
spike, now we don't.

To be accurate, there used to be a huge and unavoidable spike, now there's
a control that aims to make it smaller. The problem hasn't completely
gone away yet.

With that in mind, let me start over. Here's what's in the release notes
right now:

"Distributed checkpoints prevent I/O spikes during checkpoints"

It's good this came up, because that is factually wrong; while the average
case is much better some OS-dependant aspects of the spike (what happens
at fsync) are certainly still there. I think it's easier to rewrite this
whole thing so it's technically accurate rather than a simple fix of the
wording, something like this:

"Checkpoint writes can be spread over a longer time period to smooth the
I/O spike during each checkpoint"

It's got spread, it's got smooth, and if I could have worked "silky" in
there too I would have.

--
* Greg Smith gsmith@gregsmith.com http://www.gregsmith.com Baltimore, MD

On Dec 7, 2007 10:25 PM, Greg Smith <gsmith@gregsmith.com> wrote:

"Smoothed" makes a lot of sense for me. We used to have a checkpoint
spike, now we don't.

To be accurate, there used to be a huge and unavoidable spike, now there's
a control that aims to make it smaller. The problem hasn't completely
gone away yet.

Agreed.

"Distributed checkpoints prevent I/O spikes during checkpoints"

It's good this came up, because that is factually wrong;

Agreed.

"Checkpoint writes can be spread over a longer time period to smooth the
I/O spike during each checkpoint"

Sounds good to me.

It's got spread, it's got smooth, and if I could have worked "silky" in
there too I would have.

:)

--
Jonah H. Harris, Sr. Software Architect | phone: 732.331.1324
EnterpriseDB Corporation | fax: 732.331.1301
499 Thornall Street, 2nd Floor | jonah.harris@enterprisedb.com
Edison, NJ 08837 | http://www.enterprisedb.com/

Greg Smith wrote:

It's good this came up, because that is factually wrong; while the average
case is much better some OS-dependant aspects of the spike (what happens at
fsync) are certainly still there. I think it's easier to rewrite this
whole thing so it's technically accurate rather than a simple fix of the
wording, something like this:

"Checkpoint writes can be spread over a longer time period to smooth the
I/O spike during each checkpoint"

Thanks, I changed it to this.

--
Alvaro Herrera http://www.advogato.org/person/alvherre
"No necesitamos banderas
No reconocemos fronteras" (Jorge Gonz�lez)

All,

Just FYI, it's going to be difficult to replace the name of the feature in
the PR docs at this point; I already have 11 translations. What's *wrong*
with "Load Distributed Checkpoint", which is what we've been calling it
for 6 months?

--
--Josh

Josh Berkus
PostgreSQL @ Sun
San Francisco

Josh Berkus wrote:

All,

Just FYI, it's going to be difficult to replace the name of the feature in
the PR docs at this point; I already have 11 translations. What's *wrong*
with "Load Distributed Checkpoint", which is what we've been calling it
for 6 months?

There was nothing *wrong* with the old wording, but the new wording is
clearer? Do you disagree it is clearer? I don't think it makes sense
to keep less-clear wording just to match press release translations.

It is not like we are changing the wording 24 hours before final
release. There will perhaps be other adjustments that might be needed
for the press release. Also, the non-English press release isn't going
to match the English release notes word-for-word anyway (they aren't in
English) so is the new naming that big an issue?

I suggest you update the English press release and ask as many
translators who want to update theirs.

--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://postgres.enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +

On Wed, 12 Dec 2007, Josh Berkus wrote:

What's *wrong* with "Load Distributed Checkpoint", which is what we've
been calling it for 6 months?

One issue was that "distributed" has some association with distributed
computing, which isn't actually the case. "Spread" is also more
descriptive of what actually ended up being committed. Those are fairly
subtle wording issues that I wouldn't necessarily expect to survive
translation.

The other problem was that the original description over-sold the feature
a bit. It said "prevent I/O spikes" when it actually just reduces them.
Still possible to have a spike, it probably won't be as big though. Your
call on whether correcting that mischaracterization is worth bothering the
translators over.

--
* Greg Smith gsmith@gregsmith.com http://www.gregsmith.com Baltimore, MD

Just FYI, it's going to be difficult to replace the name of the feature in
the PR docs at this point; I already have 11 translations. What's *wrong*
with "Load Distributed Checkpoint", which is what we've been calling it
for 6 months?

Are you saying the PR was 'string freezed' before rc1? And before the actual backend? I wonder how reasonable that really is...

That said we shouldn't change things around for no reason. IKn this case I think there was good motivation.

/Magnus

Josh Berkus wrote:

All,

Just FYI, it's going to be difficult to replace the name of the feature in
the PR docs at this point; I already have 11 translations. What's *wrong*
with "Load Distributed Checkpoint", which is what we've been calling it
for 6 months?

Is a translator really not able to change 3 words in a week? Come again.

--
Alvaro Herrera Developer, http://www.PostgreSQL.org/
Hi! I'm a .signature virus!
cp me into your .signature file to help me spread!

Alvaro Herrera wrote:

Josh Berkus wrote:

All,

Just FYI, it's going to be difficult to replace the name of the feature in
the PR docs at this point; I already have 11 translations. What's *wrong*
with "Load Distributed Checkpoint", which is what we've been calling it
for 6 months?

Is a translator really not able to change 3 words in a week? Come again.

I think it is likely more about being able to reach the translators. The
more common ones such as yourself are obvious but others may not be.

Either way, I think that the change is valid and we need to do it.

Joshua D. Drake

"Joshua D. Drake" <jd@commandprompt.com> writes:

Alvaro Herrera wrote:

Is a translator really not able to change 3 words in a week? Come again.

I think it is likely more about being able to reach the translators. The more
common ones such as yourself are obvious but others may not be.

Either way, I think that the change is valid and we need to do it.

When do we normally freeze strings?

--
Gregory Stark
EnterpriseDB http://www.enterprisedb.com
Ask me about EnterpriseDB's Slony Replication support!

Greg, All:

The other problem was that the original description over-sold the feature
a bit. It said "prevent I/O spikes" when it actually just reduces them.
Still possible to have a spike, it probably won't be as big though. Your
call on whether correcting that mischaracterization is worth bothering the
translators over.

Sounds like I'd better. Sending out this afternoon.

--
Josh Berkus
PostgreSQL @ Sun
San Francisco