I have to agree with Alfred here: this does not sound like a feature,
it sounds like a horrid hack. You're giving up *all* consistency
guarantees for a performance gain that is really going to be pretty
minimal in the WAL context.
The "buffered log" still guarantees consistency within the
I am just suggesting that instead of flushing the log on every
transaction end, just do it every X seconds.
Or maybe more practical is, when the log buffer fills.
And of course during checkpoints.
Andreas
because as said, it can be any other language besides C and also
the 'AS file' is weird.
This is interesting. It allows us to control the default behavour of
"C". I would vote to default to 7.0-style when no version is used for
7.1, then default to 7.1 style in 7.2 and later. We don't
[ Charset ISO-8859-1 unsupported, converting... ]
I am just suggesting that instead of flushing the log on every
transaction end, just do it every X seconds.
Or maybe more practical is, when the log buffer fills.
And of course during checkpoints.
Log filling is too abritrary. If I
Or maybe more practical is, when the log buffer fills.
And of course during checkpoints.
Log filling is too abritrary. If I commit something, and nothing
happens for 2 hours, we should commit that transaction.
Checkpoint, as I said ?
Andreas
[ Charset ISO-8859-1 unsupported, converting... ]
Or maybe more practical is, when the log buffer fills.
And of course during checkpoints.
Log filling is too abritrary. If I commit something, and nothing
happens for 2 hours, we should commit that transaction.
Checkpoint, as I
Here's what I'm planning to do to make the world safe for backup/restore
of user-added template1 data:
1. pg_database will have two new boolean columns, dbistemplate and
dballowconn --- if dballowconn is false then no new connections to that
database will be accepted. (An even better
Tom Lane wrote:
1. pg_database will have two new boolean columns, dbistemplate and
dballowconn --- if dballowconn is false then no new connections to that
database will be accepted. (An even better solution would be to make
a "dbisreadonly" attribute, but implementing such a thing
Lamar Owen [EMAIL PROTECTED] writes:
That is a great side-effect. Now, if there were a way to initdb just
template0, leaving everything else in place, then rebuilding template1
-- of course, a similar ability is there now, but the two-stage initdb
this implies could make pg_upgrade work
At 11:48 13/11/00 -0500, Tom Lane wrote:
5. pg_dump should ignore objects with OID = lastsysoid of the target
database.
I think it should ignore objects with OID = lastsysoid of template0; then
when it does a restore, it should use 'WITH TEMPLATE template0'. At least
this should be an
I have to agree with Peter E. on this patch: it's poorly thought out.
I don't mind the idea of being able to relocate the socket file,
but the client-side interface they've chosen is silly. Having to
add another switch to every client app is not reasonable --- it's
bad enough that you had to
Philip Warner [EMAIL PROTECTED] writes:
5. pg_dump should ignore objects with OID = lastsysoid of the target
database.
I think it should ignore objects with OID = lastsysoid of template0; then
when it does a restore, it should use 'WITH TEMPLATE template0'.
Restores will be with template
Tom Lane wrote:
Lamar Owen [EMAIL PROTECTED] writes:
That is a great side-effect. Now, if there were a way to initdb just
template0, leaving everything else in place, then rebuilding template1
I'm missing something --- I don't see how this affects pg_upgrade one
way or the other, except
I have to agree with Peter E. on this patch: it's poorly thought out.
Now you tell me. :-)
I don't mind the idea of being able to relocate the socket file,
but the client-side interface they've chosen is silly. Having to
add another switch to every client app is not reasonable --- it's
* Larry Rosenman [EMAIL PROTECTED] [001112 19:20]:
Ok, You guys are probably tired of me, BUT, here is another one, that
adds the facility to set the program name used in syslog.
(this includes the other ones).
One gotcha, the parser doesn't like special characters in strings.
For
Actually, the thing doesn't even compile:
pqcomm.c: In function `StreamServerPort':
pqcomm.c:259: warning: implicit declaration of function `hstrerror'
pqcomm.c:259: `h_errno' undeclared (first use in this function)
pqcomm.c:259: (Each undeclared identifier is reported only once
pqcomm.c:259:
Looks like it needs a resolver include
LER
* Tom Lane [EMAIL PROTECTED] [001113 13:11]:
Actually, the thing doesn't even compile:
pqcomm.c: In function `StreamServerPort':
pqcomm.c:259: warning: implicit declaration of function `hstrerror'
pqcomm.c:259: `h_errno' undeclared (first use
On BSDI, hstrerror is defined in netdb.h. Do you have it anywhere? Is
that a proper function call?
Actually, the thing doesn't even compile:
pqcomm.c: In function `StreamServerPort':
pqcomm.c:259: warning: implicit declaration of function `hstrerror'
pqcomm.c:259: `h_errno' undeclared
Bruce Momjian [EMAIL PROTECTED] writes:
On BSDI, hstrerror is defined in netdb.h. Do you have it anywhere? Is
that a proper function call?
There is no hstrerror anywhere on HPUX. h_errno is defined in
netdb.h, but only with nonstandard compilation options:
#ifdef _XOPEN_SOURCE_EXTENDED
Applied.
Ok, You guys are probably tired of me, BUT, here is another one, that
adds the facility to set the program name used in syslog.
(this includes the other ones).
One gotcha, the parser doesn't like special characters in strings.
For example, i tried to use pg-test, and if failed
OK, working on it now.
Thanks.
You need to remove it from src/backend/postmaster/postmaster.c as
well..
cc -O -K inline -I/usr/local/include -I../../../../src/include -c -o guc.o guc.c
cc -O -K inline -I/usr/local/include -I../../../../src/include -c -o guc-file.o
guc-file.c
cc
At 12:38 13/11/00 -0500, Tom Lane wrote:
I think it should ignore objects with OID = lastsysoid of template0; then
when it does a restore, it should use 'WITH TEMPLATE template0'.
Restores will be with template template0 anyway, but it seems to me that
dump should look to lastsysoid of the
Larry Rosenman writes:
Ok, You guys are probably tired of me, BUT, here is another one, that
adds the facility to set the program name used in syslog.
(this includes the other ones).
Why would one want to do that? ISTM that that would just be a means to
play games with the
Bruce Momjian [EMAIL PROTECTED] writes:
+ if (conn-pghostaddr conn-pghostaddr[0] == '/')
+ {
+ if (conn-pgunixsocket)
+ free(conn-pgunixsocket);
+ conn-pgunixsocket = conn-pghostaddr;
+ conn-pghostaddr = NULL;
+ }
I
Peter Eisentraut [EMAIL PROTECTED] writes:
What happened to the concerns that were raised? The socket file is a lock
file, you cannot just move it around.
Good point. IIRC, we rely on the socket file lock to ensure that you
can't start two postmasters with the same port number. (If both are
On Saturday 11 November 2000 20:30, Tom Lane wrote:
Philip Warner [EMAIL PROTECTED] writes:
Might even be able to implement a 'CREATE MODULE' which loads an
object, enquires about functions in the module, and creates function
entries based on information returned from pg_fmgr_info...
At 20:33 13/11/00 -0500, Mark Hollomon wrote:
On Saturday 11 November 2000 20:30, Tom Lane wrote:
Philip Warner [EMAIL PROTECTED] writes:
Might even be able to implement a 'CREATE MODULE' which loads an
object, enquires about functions in the module, and creates function
entries based on
I've got the multiple-template-database feature nearly ready to commit,
but I was just noticing that the way we handle default MB encoding
selection seems wrong for this way of looking at things. Currently,
if you don't specify "ENCODING = something" in CREATE DATABASE, the
default is to use the
[Cced to hackers list]
BTW I have found another bug with EUC_TW support. line 917 in conv.c:
*p++ = c1 - LC_CNS11643_3 + 0xa3;
this should be:
*p++ = *mic++ - LC_CNS11643_3 + 0xa3;
Otherwise, CNS 11643-1992 Plane 3 or more won't
Hiroshi Inoue [EMAIL PROTECTED] writes:
I fixed a few bugs caused by SearchSysCacheTuple()
recently. There seems to remain more similar bugs.
Yes, I'm sure there are a lot :-(. We have talked about solving this
by using some form of reference-counting in the cache, but I don't
believe
fixed, let mek now if there are any others I've missed ...
On Mon, 13 Nov 2000, Larry Rosenman wrote:
Are the @postgresql.org addresses supposed to work?
-Original Message-
From: Mail Delivery Subsystem [mailto:[EMAIL PROTECTED]]
Sent: Monday, November 13, 2000 4:13 PM
To:
Tom Lane wrote:
Hiroshi Inoue [EMAIL PROTECTED] writes:
I fixed a few bugs caused by SearchSysCacheTuple()
recently. There seems to remain more similar bugs.
Yes, I'm sure there are a lot :-(. We have talked about solving this
by using some form of reference-counting in the cache, but I
Hiroshi Inoue [EMAIL PROTECTED] writes:
Isn't it practical to replace all susipicious Search
SysCacheTuple() by SearchSysCacheTupleCopy() ?
That would replace a rare failure condition by a not-at-all-rare
memory leak. I'm not sure there'd be a net gain in reliability :-(
A more serious
BTW, I thought you were backing out the unnecessary changes to the
client applications? pg_dump seems not to be reverted yet, for one...
I thought I had. I don't see them here. Can you tell me what you see.
--
Bruce Momjian| http://candle.pha.pa.us
[EMAIL
Bruce Momjian [EMAIL PROTECTED] writes:
BTW, I thought you were backing out the unnecessary changes to the
client applications? pg_dump seems not to be reverted yet, for one...
I thought I had. I don't see them here. Can you tell me what you see.
My apologies. I must have been looking at
Bruce Momjian [EMAIL PROTECTED] writes:
+ if (conn-pghostaddr conn-pghostaddr[0] == '/')
+ {
+ if (conn-pgunixsocket)
+ free(conn-pgunixsocket);
+ conn-pgunixsocket = conn-pghostaddr;
+ conn-pghostaddr = NULL;
+ }
I would be
Peter Eisentraut [EMAIL PROTECTED] writes:
What happened to the concerns that were raised? The socket file is a lock
file, you cannot just move it around.
Good point. IIRC, we rely on the socket file lock to ensure that you
can't start two postmasters with the same port number. (If
Bruce Momjian [EMAIL PROTECTED] writes:
I think we had some discussions about changing the way that shared
memory keys are generated, which might make this a less critical issue.
But until something's done about that, this patch looks awfully
dangerous.
But do we yank it out for that
Bruce Momjian [EMAIL PROTECTED] writes:
I think we had some discussions about changing the way that shared
memory keys are generated, which might make this a less critical issue.
But until something's done about that, this patch looks awfully
dangerous.
But do we yank it out for that
BTW, I thought you were backing out the unnecessary changes to the
client applications? pg_dump seems not to be reverted yet, for one...
regards, tom lane
Bruce Momjian writes:
Bruce Momjian writes:
I am tempted to apply this. This is the second person who asked for
binding to a single port. The patch looks quite complete, with doc
changes. It appears to be a thorough job.
Postmaster options are evil, please put
OK, I have removed the -k unix socketpath option from the client side of
this patch, and modified libpq so if they specify a host with a leading
slash, it will be considered a unix socket path. Attached is the
relevant patch to libpq.
Bruce Momjian writes:
Bruce Momjian writes:
* Peter Eisentraut [EMAIL PROTECTED] [001113 17:41]:
Bruce Momjian writes:
Applied.
Uh, shouldn't the problems Larry pointed out be fixed first?
I don't think the problems are BECAUSE of my code. The new stuff
works, just putting the new options in postgresql.conf with a bad
* Peter Eisentraut [EMAIL PROTECTED] [001113 17:41]:
Bruce Momjian writes:
Applied.
Uh, shouldn't the problems Larry pointed out be fixed first?
I don't think the problems are BECAUSE of my code. The new stuff
works, just putting the new options in postgresql.conf with a bad
value causes
* Peter Eisentraut [EMAIL PROTECTED] [001113 17:43]:
Larry Rosenman writes:
Ok, You guys are probably tired of me, BUT, here is another one, that
adds the facility to set the program name used in syslog.
(this includes the other ones).
Why would one want to do that? ISTM that that
There were problems? I saw someone asking why it hadn't been applied
yet. What do people want? I can back it out.
Bruce Momjian writes:
Applied.
Uh, shouldn't the problems Larry pointed out be fixed first?
Ok, You guys are probably tired of me, BUT, here is another one,
Larry Rosenman writes:
* Peter Eisentraut [EMAIL PROTECTED] [001113 17:43]:
Larry Rosenman writes:
Ok, You guys are probably tired of me, BUT, here is another one, that
adds the facility to set the program name used in syslog.
(this includes the other ones).
Why would one
Tom Lane writes:
I've got the multiple-template-database feature nearly ready to commit,
but I was just noticing that the way we handle default MB encoding
selection seems wrong for this way of looking at things. Currently,
if you don't specify "ENCODING = something" in CREATE DATABASE, the
And now it's on www.slashdot.org ...
http://slashdot.org/articles/00/11/13/1342208.shtml
Poul L. Christiansen
Michael Fork wrote:
Thought this may be of interest to some...
http://www.phpbuilder.com/columns/tim20001112.php3
Michael Fork - CCNA - MCP - A+
Network Support - Toledo
[ Charset ISO-8859-1 unsupported, converting... ]
And now it's on www.slashdot.org ...
http://slashdot.org/articles/00/11/13/1342208.shtml
Poul L. Christiansen
Michael Fork wrote:
Thought this may be of interest to some...
http://www.phpbuilder.com/columns/tim20001112.php3
And now it's on www.slashdot.org ...
http://slashdot.org/articles/00/11/13/1342208.shtml
Poul L. Christiansen
Michael Fork wrote:
Thought this may be of interest to some...
http://www.phpbuilder.com/columns/tim20001112.php3
Michael Fork - CCNA - MCP - A+
Network Support - Toledo
I made it all the way through the article. I'll summarize it for you:
Postgres - hooray!
MySQL - boo!
Yeah, and that's about it. No analysis or anything. Disappointing, after
waiting so long for the pages to load.
Since this is an open source database article linked off of slashdot, I
imagine
* Bruce Momjian [EMAIL PROTECTED] [001113 15:36]:
Applied.
Thanks. FWIW, I had to pull the CVS copy of guc.c (my cvs update got
two copies of my updates to the table).
Comments on my postgresql.conf issues from the rest of the -hackers
corp?
Thanks!
--
Larry Rosenman
You need to remove it from src/backend/postmaster/postmaster.c as
well..
cc -O -K inline -I/usr/local/include -I../../../../src/include -c -o guc.o guc.c
cc -O -K inline -I/usr/local/include -I../../../../src/include -c -o guc-file.o
guc-file.c
cc -O -K inline -I/usr/local/include
54 matches
Mail list logo
