rade started creating an incremental-analyze script, people
thought it should be more generic so it was moved to vacuumdb
--analyze-in-stages. Seems we should do the same thing for the
extension upgrade script.
--
Bruce Momjian http://momjian.us
EnterpriseDB
tion, but two different TLS
> connections, and make the authentication fail.
>
> SSL certificates, with validation, achieves the same, but channel binding
> achieves it without the hassle of certificates.
How does it do that?
--
Bruce Momjian http://momjian.us
EnterpriseDB
ink?
Well, we have been using JavaScript for years, so my patch only fixes
the JavaScript we already have. Seems applying the fix now makes sense
and then we can ponder other methods.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
(backups which include pg_authid) is just a "check-box" item.
Well, they have the entire backup so I don't think cracking the password
is a huge win, though the password does potentially give them access to
future data. And it prevents them from reusing the password on anoth
On Tue, Apr 11, 2017 at 04:51:03PM -0400, Tom Lane wrote:
> Bruce Momjian writes:
> Yeah, very long ago. A quick search of our archives shows that the number
> of mentions of Borland pretty much fell off a cliff after 2009 (excluding
> the repeated conversations about dropping supp
And apologies
> > for keep pushing it till the last minute even though it was clear to me
> > quite some time back the patch is not going to make it.
>
> What confuses me about that position is that people were advocating to
> actually commit till literally hou
e big win for SCRAM is the inability to replay md5 packets
after recording 16k sessions (salt was only 32-bit, so a 50% chance of
replay after 16 sessions), and storage of SHA256 hashes instead of MD5
in pg_authid, though the value of that is mostly a check-box item
because collisions are not a pr
;
>
> Ok, since we have two votes for it, I will go ahead and do that (as a separate
> patch pushed together).'
I am fine with removing things, but I do remember one reason the Borland
part was kept is that some tool would only work with the
Borland-compiled library, not gcc or MSVC,
> SCRAM-SHA-512 verifier, not both, so that will dictate which one to use.
It seems openssl has had to deal with cipher preferences for years and
invented ssl_ciphers.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you ar
rth the trouble.)
We already have code to do all of that, but it was removed from
pg_upgrade in 9.5. You can still see it in 9.4:
contrib/pg_upgrade/version_old_8_3.c::old_8_3_invalidate_hash_gin_indexes()
I would be happy to restore that code and make it w
l as that is. We now have
something to look forward to in Postgres 11. :-)
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be. +
+ Ancient Roman grave inscription
er on
something the build system does anyway so that if someone builds pre-10
docs with the new build system, they get the same output. The attached
patch uses:
div.refnamediv h2 .refentrytitle
and "div.refnamediv" is only generated in the new build system.
--
Bruce Momjian
On Sun, Apr 9, 2017 at 06:59:09PM +0200, Magnus Hagander wrote:
> On Sat, Apr 8, 2017 at 3:52 AM, Bruce Momjian wrote:
> Attached is a patch that can be applied to pgweb which should fix all of
> this.
>
> Is there any chance we can find a way to do this with actual CS
On Sun, Apr 9, 2017 at 07:00:38PM +0200, Magnus Hagander wrote:
>
>
> On Sun, Apr 9, 2017 at 2:32 AM, Bruce Momjian wrote:
>
> On Sat, AprĀ 8, 2017 at 12:50:19PM -0400, Robert Haas wrote:
> > On Sat, Apr 8, 2017 at 6:39 AM, Bruce Momjian wrote:
> >
On Sat, Apr 8, 2017 at 12:50:19PM -0400, Robert Haas wrote:
> On Sat, Apr 8, 2017 at 6:39 AM, Bruce Momjian wrote:
> > What other problems do we have with pgweb that I can work on?
>
> Well, the 10devel documentation doesn't believe in orange. Compare:
>
> https
doc looks ok now.
Great.
What other problems do we have with pgweb that I can work on?
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be. +
+ Ancient Roman grave inscrip
ied, is to conditionally inject additional CSS
to be _more_ specific than the first CSS and set the font-size to a
simple '1em' so the first CSS is not called twice. I don't think
'important!' is necessary but it would be good to test this.
Attached is a patch that can be a
t; PL/Java ordinarily checks what version of its schema is around, but
> that may be premature while pg_upgrade is doing its thing and the schema
> might not be all there yet.
You can check the backend global variable IsBinaryUpgrade to check if
binary upgrade is being performed. Does that hel
On Thu, Mar 30, 2017 at 06:23:04PM -0400, Robert Haas wrote:
> Hi,
>
> One of the things we'll need to do to get 10beta1 out the door is have
> a set of release notes. I spoke with Bruce Momjian today at PGCONF.US
> and he told me that he has set aside time during April to w
On Tue, Mar 21, 2017 at 04:04:58PM -0400, Bruce Momjian wrote:
> On Tue, Mar 21, 2017 at 04:56:16PM -0300, Alvaro Herrera wrote:
> > Bruce Momjian wrote:
> > > On Tue, Mar 21, 2017 at 04:43:58PM -0300, Alvaro Herrera wrote:
> > > > Bruce Momjian wrote:
> > &
conference attendance. If that delays our final release for a
week, it is worth it.
I propose we go for a week delay in closing the commit fest, and we
decide right now. Ideally I like to to see delay in one-week increments
_and_ announce that a week before each deadline.
--
Bruce Momjian
;
> Let me know if its still not clear and I can draw some diagrams to explain it.
Ah, yes, that does help to explain the 50% because 50% of updates are
now HOT/WARM.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterpris
g entry, not exactly how
> it's spelled.
Yes, it doesn't matter as long as it is stated somehow. I don't know of
any missing cases due to text differences.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As y
would hope that most updates get done within a single
> block, so that the row-moves-to-new-block case wouldn't happen very
> often.
>
> I'm perplexed.
Yes, I asked the same question in this email:
https://www.postgresql.org/message-id/2017032119.ge16...@momjian.us
y default.
Maybe its time to have a documentation section listing suggested changes
for small installs so we can have more reasonable defaults.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As
On Tue, Mar 21, 2017 at 04:56:16PM -0300, Alvaro Herrera wrote:
> Bruce Momjian wrote:
> > On Tue, Mar 21, 2017 at 04:43:58PM -0300, Alvaro Herrera wrote:
> > > Bruce Momjian wrote:
> > >
> > > > I don't think it makes sense to try and save bits and ad
On Tue, Mar 21, 2017 at 04:43:58PM -0300, Alvaro Herrera wrote:
> Bruce Momjian wrote:
>
> > I don't think it makes sense to try and save bits and add complexity
> > when we have no idea if we will ever use them,
>
> If we find ourselves in dire need of addi
nd then HEAP_WARM_UPDATED just from correctness
> perspective.
I don't think it makes sense to try and save bits and add complexity
when we have no idea if we will ever use them, but again, I am back to
my original question of whether we have done sufficient research, and if
everyone says "
zy
vacuums would show no improvementin in real-world workloads, but maybe I
am wrong.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be. +
+ Ancient Roman grave
many bits we need, and my concern is that we
haven't discussed the improvement ideas enough to know we have done the
best we can for PG 10.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I a
it's going to be really painful for a lot
> of users. Many users can't realistically upgrade using pg_dump, ever.
> So they'll be stuck on the release before the one that breaks
> compatibility for a very long time.
Right. If we weren't setting tuple and tid bits
r. I guess I just don't feel we have explored the
duplicate-index-entry problem enough for me to be comfortable.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be
e with it, I am too, but I have to ask these questions.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be. +
+ Ancient Roman grave inscription +
--
Sent via pg
On Mon, Mar 20, 2017 at 11:57:13AM -0400, Stephen Frost wrote:
> Tom,
>
> * Tom Lane (t...@sss.pgh.pa.us) wrote:
> > Stephen Frost writes:
> > > * Bruce Momjian (br...@momjian.us) wrote:
> > >> 1. make the change now and mention it in the release notes
>
On Mon, Mar 20, 2017 at 11:40:34AM -0400, Tom Lane wrote:
> Stephen Frost writes:
> > * Bruce Momjian (br...@momjian.us) wrote:
> >> 1. make the change now and mention it in the release notes
> >> 2. #1, but also provide backward compatibility for 5+ years
&
is for changes
that could affect data validity or are hard to detect, e.g. the
standard-conforming string change.
Is this accurate? If so we can write this down and avoid the +1/-1
re-litigating every time we want to change something, or at least have a
more focused conversation on the issues
when do you sleep?
>
> I think that applies to a bunch of people, including yourself ;)
Gee, no one asks when I sleep. I wonder why. ;-)
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you ar
for PG 10:
commit 10c064ce4dad088ba2d8b978bff6009b9f22dc3a
Author: Bruce Momjian
Date: Tue Oct 25 11:26:15 2016 -0400
Consistently mention 'SELECT pg_reload_conf()' in config files
Previously we only mentioned SIGHUP and 'pg_ctl reload'
p track of the hash
> separately to the password text and use prefixes instead, and since we
> need compatibility for dumps, it's hard to do a great deal about
> though.
With SCRAM coming in PG 10, is there anything we can do to clean this up
for PG 10?
--
Bruce Momjian ht
ound protections
> are disabled because oldest checkpointed MultiXact %u does not exist on
> disk". Pushed without that.
Gee, I kind of like the new messages:
LOG: listening on IPv4 address "127.0.0.1", port 5432
LOG: listening on Unix socket "/tmp/.s.PGS
instead of 's' is that the latter was
> taken for "special" relations, which we removed a few releases ago
> (commit 3a694bb0a1).
Ah! I knew there was a reason for 'S', but couldn't remember it. :-)
--
Bruce Momjian http://momjian.us
Enterpr
only downside of having it in 10.0
is that it will not have lived in the source tree for as long a time
between commit and PG release, but if it is tested, that seems fine.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are,
On Fri, Mar 10, 2017 at 11:23:54AM +0100, Andreas Karlsson wrote:
> On 03/09/2017 11:25 PM, Bruce Momjian wrote:
> >"data" and "base" where chosen because it is a "data-base", but with the
> >pg_ prefixes it would be a pg_data_pg_base. ;-)
>
&
look in
pg_stat_statements to see how long is usually ran. It seemed like a
useful idea to me.
I don't think they can hash the query manually because of the constants
involved.
--
Bruce Momjian http://momjian.us
EnterpriseDB http
care much about which name
> is picked as long as we get rid off the "pg_" prefix.
>
> Btw, is there a reason for why global and base do not have the "pg_" prefix?
"data" and "base" where chosen because it is a "data-base", but with
ider disallowing multiple queries in PQexec()
as an additional barrier to SQL injection attacks
and it is a "consider" item. Should it be moved to the Wire Protocol
Changes / v4 Protocol section or removed?
--
Bruce Momjian http://momjian.us
EnterpriseDB
7;r', 'm') AND "
but partitions are defined as 'P':
#define RELKIND_PARTITIONED_TABLE 'P' /* partitioned table */
I am a little confused by the above. Is a partitioned table the parent
or the children? Reading the
t, but "log" seems perhaps a little too
> generic to me. Would something like "server_log" be better?
"activity_log"? I like the idea of a rename.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
e64 is less easy to compress because
input bytes span base64-bytes, so "dog" might encode differently
depending on where the two high bits are stored, while hex alway encodes
"dog" the same way.
--
Bruce Momjian http://momjian.us
EnterpriseDB
y the best you can do. Your analysis of how we
used to treat large objects is correct, and was never adjusted for the
changes you outlined.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am
her place by
pre-9.0
* VACUUM FULL; kept for binary
* upgrade support */
However, if users built Postgres 8.4 with integer timestamps, they could
still be in the data files. Does a
ndering if people are fine
> > to include it in this last CF.
>
> The March CF is already looking pretty daunting. We can try to include
> this but I won't be too surprised if it gets punted to a future CF.
Yeah, that was my reaction too.
--
Bruce Mom
On Mon, Feb 27, 2017 at 01:00:20PM -0800, David E. Wheeler wrote:
> On Feb 27, 2017, at 12:04 PM, Bruce Momjian wrote:
>
> > Just stating the obvious, but one of the reasons CPAN works so well is
> > that most of the modules are written in Perl and hence don't need
>
s
> not a patch in the easy category, far from that, but it does not
> present a new concept.
FYI, the thread started on 2013-11-15.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you
to what we deliver in core.
Just stating the obvious, but one of the reasons CPAN works so well is
that most of the modules are written in Perl and hence don't need
per-platform compilation.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterp
reindex (they throw a meaningful error
> already due to the lack of indexes.)
I don't think we are doing this, but if the parent table doesn't have a
physical file pg_upgrade will need to be taught that. We have that case
now for unlogged tables on standby servers that we need to addr
On Sat, Feb 25, 2017 at 06:49:01PM -0500, Peter Eisentraut wrote:
> On 2/23/17 17:55, Bruce Momjian wrote:
> > On Thu, Feb 23, 2017 at 04:08:58PM -0500, Tom Lane wrote:
> >> Bruce Momjian writes:
> >>> Is there a reason we don't support base64 as a bytea_output
On Wed, Feb 22, 2017 at 07:44:41AM +0530, Robert Haas wrote:
> On Tue, Feb 21, 2017 at 2:03 AM, Bruce Momjian wrote:
> > I have to admit my reaction was similar to Simon's, meaning that the
> > lack of docs is a problem, and that the limitations are kind of a
> > surpris
this standby to rejoin the primary, you just
want to upgrade it.
Second, I thought unlogged tables had empty files on the standby, not
_missing_ files. Is that correct? Should pg_upgrade just allow missing
unlogged table files? I don't see any way to detect we are running on a
standby since th
eral acceptance to that idea..
> I
> am assuming latter :-))
Yes, I think it is the latter.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be. +
+ Ancient
having two build systems we have to maintain
is better than 1.1 build systems where we can mostly ignore 0.1 of that.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be. +
+
On Fri, Feb 24, 2017 at 10:09:50PM +0200, Ants Aasma wrote:
> On Fri, Feb 24, 2017 at 9:37 PM, Bruce Momjian wrote:
> > Oh, that's why we will hopefully eventually change the page checksum
> > algorithm to use the special CRC32 instruction, and set a new checksum
> > ver
a way they are a GPU built
into CPUs.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be. +
+ Ancient Roman grave inscription +
--
Sent via pgsql-hackers maili
never got around to adding runtime checks for
> the architecture to enable this speedup.
Oh, that's why we will hopefully eventually change the page checksum
algorithm to use the special CRC32 instruction, and set a new checksum
version --- got it. I assume there is currently no compile-t
e. IIRC it was actually a lot faster than the naive hardware
> CRC that is used for WAL and about on par with interleaved CRC.
Uh, I thought already did compile-time testing for SSE4.1 and used them
if present. Why do you say "with default compiles there is a lot of
performance left
On Fri, Feb 24, 2017 at 02:14:23PM +0530, Pavan Deolasee wrote:
>
>
> On Thu, Feb 23, 2017 at 11:53 PM, Bruce Momjian wrote:
>
> On Thu, Feb 23, 2017 at 03:03:39PM -0300, Alvaro Herrera wrote:
> > Bruce Momjian wrote:
> >
> > > As I remember,
On Thu, Feb 23, 2017 at 07:09:57PM -0500, Andrew Dunstan wrote:
>
>
> On 02/23/2017 06:52 PM, David Fetter wrote:
> > On Thu, Feb 23, 2017 at 05:55:37PM -0500, Bruce Momjian wrote:
> >> On Thu, Feb 23, 2017 at 04:08:58PM -0500, Tom Lane wrote:
> >>> Bruc
On Thu, Feb 23, 2017 at 04:08:58PM -0500, Tom Lane wrote:
> Bruce Momjian writes:
> > Is there a reason we don't support base64 as a bytea_output output
> > option, except that no one has implemented it?
>
> How about "we already have one too many bytea output fo
ze the risk of losing data.
Do these other databases do checksums because they don't do
full_page_writes? They just detect torn pages rather than repair them
like we do?
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprise
https://en.wikipedia.org/wiki/Base64
Is there a reason we don't support base64 as a bytea_output output
option, except that no one has implemented it?
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so o
On Thu, Feb 23, 2017 at 03:58:59PM -0300, Alvaro Herrera wrote:
> Bruce Momjian wrote:
> > On Thu, Feb 23, 2017 at 03:45:24PM -0300, Alvaro Herrera wrote:
>
> > > > and potentially trim the first HOT chain as those tuples become
> > > > invisible.
> > >
On Thu, Feb 23, 2017 at 03:45:24PM -0300, Alvaro Herrera wrote:
> Bruce Momjian wrote:
>
> > Well, let's walk through this. Let's suppose you have three updates
> > that stay on the same page and don't update any indexed columns --- that
> > would produce a
On Thu, Feb 23, 2017 at 03:26:09PM -0300, Alvaro Herrera wrote:
> Bruce Momjian wrote:
> > On Thu, Feb 23, 2017 at 03:03:39PM -0300, Alvaro Herrera wrote:
> > > Bruce Momjian wrote:
> > >
> > > > As I remember, WARM only allows
> > > > a si
On Thu, Feb 23, 2017 at 03:03:39PM -0300, Alvaro Herrera wrote:
> Bruce Momjian wrote:
>
> > As I remember, WARM only allows
> > a single index-column change in the chain. Why are you seeing such a
> > large performance improvement? I would have thought it would be tha
m
> block size. Can look at that if it comes to finding more bits.
OK, so it seems more bits is not a blocker to enhancements, yet.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you w
it has found an existing bug and given good performance
numbers. :-)
I have what might be a supid question. As I remember, WARM only allows
a single index-column change in the chain. Why are you seeing such a
large performance improvement? I would have thought it would be that
high if we allowed
: 'worktree' is not a git command. See 'git --help'.
which is in Debian Jessie. This reports worktree was added in 2.5,
released in July 2015:
https://developer.atlassian.com/blog/2015/10/cool-features-git-2.x/
I am glad the git team has seen the value in work tre
would be
useless for 99% of users. In talking to Alexander Korotkov, he
mentioned that indirect indexes could be used for global/cross-partition
indexes, and for index-organized tables (heap and index together in a
single file). This would greatly expand the usefulness of indirect
indexes and
logs/pgblog/2012.html#April_23_2012
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be. +
+ Ancient Roman grave inscription +
--
Sent via pgsql-hackers mailin
ally, pg_upgrade already checks for some odd directories stored
inside of PGDATA:
WARNING: new data directory should not be inside the
old data directory, e.g. %s\n", old_cluster_pgdata);
WARNING: user-defined tablespace locations should
not be inside the data d
On Tue, Feb 21, 2017 at 01:27:53AM +0100, Tomas Vondra wrote:
> On 02/21/2017 12:13 AM, Bruce Momjian wrote:
> >At the risk of asking a stupid question, we already have optimizer
> >statistics on expression indexes. In what sense are we using this for
> >multi-variate st
like the fact
we can detect collation changes with ICU, we don't know if ICU
collations change more often than OS collations.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you wi
On Mon, Feb 20, 2017 at 02:54:22PM -0800, Peter Geoghegan wrote:
> On Mon, Feb 20, 2017 at 2:38 PM, Bruce Momjian wrote:
> > I can't think of any cases where we warn of possible corruption only in
> > the server logs.
>
> It's just like any case where there w
blogs/pgblog/2017.html#February_20_2017
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be. +
+ Ancient Roman grave inscription +
--
Sent via pgsql-hackers ma
on.
> http://www.postgresql.org/mailpref/pgsql-hackers
I can't think of any cases where we warn of possible corruption only in
the server logs.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I
king them public.
It is often not clear what decisions were made and why. The idea that
unique indexes on a parent table can't guarantee uniqueness across child
tables is both a surprise, and obvious once stated.
--
Bruce Momjian http://momjian.us
EnterpriseDB
pulate the sequence and the
DEFAULT separately, but if you drop the table the sequence is dropped to
automatically.
This seems like an instructive example of how we have bundled behavior
together in the past in a logical and easy-to-understand way. Of
course, their might be some technical limit
On Thu, Feb 2, 2017 at 07:18:45AM -0800, David Fetter wrote:
> On Thu, Feb 02, 2017 at 03:16:29PM +0000, Bruce Momjian wrote:
> > I just don't see this patch going in. I think it needs are larger
> > approach to the problems it is trying to solve. I think it then
>
I just don't see this patch going in. I think it needs are larger approach to
the problems it is trying to solve. I think it then will be very useful.
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/p
fast"
> argument from this function altogether.
Agreed, it should be remove. Should I do it?
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be. +
+ Ancient
On Wed, Jan 11, 2017 at 03:08:58PM -0500, Bruce Momjian wrote:
> On Wed, Jan 11, 2017 at 08:56:23PM +0100, Fabien COELHO wrote:
> >
> > >I think we need to focus on things that _can't_ be done first, rather
> > >than things that require porting, e.g. until we h
uot;
complaints we used to get, perhaps related to exceptions. Once we had
them, the complaints of that type disappeared.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be. +
+
s, you
couldn't migrate an application that needed it. It wasn't a question of
porting --- there was just no way to port it.
Those _missing_ pieces should be a priority.
--
Bruce Momjian http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you
al: schema PL session variables
proposal: session server side variables
Is there anything that needs updating there, or it is just a question of
getting someone to implement it?
--
Bruce Momjian http://momjian.us
EnterpriseDB
On Wed, Jan 11, 2017 at 12:24:55PM -0300, Alvaro Herrera wrote:
> Bruce Momjian wrote:
>
> > Therefore, I think we need WARM done first, then we can test indirect
> > indexes to see if they are a sufficient win to add it for the small
> > percentage of users who will use i
On Wed, Jan 11, 2017 at 08:28:28AM -0500, Andrew Dunstan wrote:
>
>
> On 01/10/2017 09:25 PM, Bruce Momjian wrote:
> >I am not saying we shouldn't do it, but I am afraid that the complexity
> >in figuring out when to use indirect indexes, combined with the number
>
ances we should
recommend the user of the feature.
However, I am concerned that doing the two features serially (not in
parallel) might mean that the second feature doesn't make it into
Postgres 10, but considering we will live with this feature probably
forever, I think it is the best course
On Tue, Jan 10, 2017 at 09:25:05PM -0500, Bruce Momjian wrote:
> Thank you for the summary. I think we have to consider two things with
> indirect indexes:
>
> 1. What percentage speedup is the _average_ user going to get? You
> have to consider people who will use indirect ind
On Tue, Jan 10, 2017 at 11:36:24PM -0300, Alvaro Herrera wrote:
> Bruce Momjian wrote:
>
> > 1. What percentage speedup is the _average_ user going to get? You
> > have to consider people who will use indirect indexes who get no benefit
> > or a net slowdown, and user
figuring out when to use indirect indexes, combined with the number
of users who will try them, really hurts its inclusion.
FYI, we have a similar issue in adding GUC variables, which I outlined
in this blog post:
http://momjian.us/main/blogs/pgblog/2009.html#January_10
201 - 300 of 17339 matches
Mail list logo