Tom Lane wrote:
Pavan Deolasee [EMAIL PROTECTED] writes:
Good point. I'm envisioning a procarray.c function along the
lines of
bool TransactionHasSnapshot(xid)
which returns true if the xid is currently listed in PGPROC
and has a nonzero xmin. CIC's cleanup wait loop would check
Hi,
As a part of my university project, I am trying to modify the postgres code
to support parallel system.
As the first step I have partitioned the data on different processors. And
have kept a master node to process all the query requests. Whenever my
master node is queried I need to push my
Tom Lane wrote:
Andrew Dunstan [EMAIL PROTECTED] writes:
How would we do that? Not create the array types in bootstrap mode? Or
just special-case pg_statistic?
Not generate them in bootstrap mode works for me. IIRC, there's code
somewhere in there that allows anyarray to pass as a
Andrew Dunstan [EMAIL PROTECTED] writes:
Regarding catalog objects, we might have to try a little harder than
just not generating in bootstrap mode - IIRC we generate system views
(including pg_stats) in non-bootstrap mode. Maybe we just need to exempt
anything in the pg_catalog namespace.
Pavan Deolasee [EMAIL PROTECTED] writes:
If you haven't finished this yet, would you like me to work
on this ? If I do it, I would mostly follow the path you
suggested above, unless I run into something else.
I'm not intending to work on it.
regards, tom lane
Shaunak Godbole [EMAIL PROTECTED] writes:
I there a way of accessing other backends through the master backend.
It seems you're trying to reinvent contrib/dblink.
regards, tom lane
---(end of broadcast)---
TIP 7: You can
Thinking about XPath's output in cases such as 'SELECT xpath('/a', 'b
/');' I've realized that in such cases an empty array should be
returned (now we have NULL for such cases).
Why? Because database _knows_ that there is no element -- this is not
NULL's case (unknown).
Then I've examined how
Nikolay Samokhvalov [EMAIL PROTECTED] writes:
As I can see here, when I ask for element that doesn't exist, the
database returns NULL for me. Maybe it's well-known issue (and
actually I understood this behaviour before), but strictly speaking it
seems wrong for me: the database _knows_ that
Historically, pg_filedump
http://sources.redhat.com/rhdb/utilities.html
has relied on the size of a page's special space to determine which kind
of index it is looking at (btree, hash, etc) so that it can dump the
contents of the special space nicely. This is pretty ugly of course,
but there
Ühel kenal päeval, E, 2007-04-09 kell 10:56, kirjutas Tom Lane:
Shaunak Godbole [EMAIL PROTECTED] writes:
I there a way of accessing other backends through the master backend.
It seems you're trying to reinvent contrib/dblink.
Or you may want to use pl/proxy
(
The other platform I've whined about missing for some time is HP-UX,
especially on PA-RISC. But that's a whole different story.
there are more obscure and rare platforms(both in terms that might be a
win for the buildfarm but HP-UX is really missing.
Hello,
I have access to a PA-RISC
Nikolay Samokhvalov wrote:
2. what should I do with XPath function? There is strong analogy
between its case and array's case in my mind... Should I leave NULLs,
or empty arrays are better?
Empty array appears to be correct. The fact that arrays don't appear to
work as you might like is a
On Mon, 9 Apr 2007, Adrian Maier wrote:
The other platform I've whined about missing for some time is HP-UX,
especially on PA-RISC. But that's a whole different story.
there are more obscure and rare platforms(both in terms that might be a
win for the buildfarm but HP-UX is really missing.
On Mon, Apr 09, 2007 at 10:40:49AM -0400, Tom Lane wrote:
Andrew Dunstan [EMAIL PROTECTED] writes:
Regarding catalog objects, we might have to try a little harder than
just not generating in bootstrap mode - IIRC we generate system views
(including pg_stats) in non-bootstrap mode. Maybe
Adrian Maier wrote:
I have access to a PA-RISC machine running HP-UX 11.11. Unfortunately
the machine is on a dedicated network and has no Internet access.
It should be possible to create a mirror of the CVS repository on my
machine
(which has access to both the Internet and the dedicated
Alvaro Herrera wrote:
Thanks! I had already incorporated the foreach_worker changes into my
code, and later realized that there's an important bug regarding the
PGPROC of the workers, so I've reworked the patch, which meant that the
foreach_worker() macro went away completely.
FWIW, the
Larry Rosenman ler@lerctr.org writes:
I think I'll be able to set up my HP-UX 11.11 box here, as soon as it gets
fixed, and assuming either the bundled compiler will work or I can get
GCC on it.
If the bundled compiler is still the same non-ANSI-C weakling that was
bundled in HPUX 10, there's
Tom Lane wrote:
Stuart Bishop [EMAIL PROTECTED] writes:
After a test is run, the test harness kills any outstanding connections so
we can drop the test database. Without this, a failing test could leave open
connections dangling causing the drop database to block.
Just to make it perfectly
Tom Lane wrote:
Stuart Bishop [EMAIL PROTECTED] writes:
After a test is run, the test harness kills any outstanding connections so
we can drop the test database. Without this, a failing test could leave open
connections dangling causing the drop database to block.
Just to make it
On Mon, 9 Apr 2007, Tom Lane wrote:
Larry Rosenman ler@lerctr.org writes:
I think I'll be able to set up my HP-UX 11.11 box here, as soon as it gets
fixed, and assuming either the bundled compiler will work or I can get
GCC on it.
If the bundled compiler is still the same non-ANSI-C weakling
On Mon, Apr 09, 2007 at 10:14:41AM -0400, Andrew Dunstan wrote:
. defer for the present any consideration of a CREATE TYPE foo AS ARRAY
... command.
What is the rationale for allowing people to name the array type. When
I originally proposed the syntax I presumed that the array name would
be
Martijn van Oosterhout wrote:
On Mon, Apr 09, 2007 at 10:14:41AM -0400, Andrew Dunstan wrote:
. defer for the present any consideration of a CREATE TYPE foo AS ARRAY
... command.
What is the rationale for allowing people to name the array type. When
I originally proposed the syntax I
On Mar 25, 2007, at 12:31 PM, Tom Lane wrote:
Jim Nasby [EMAIL PROTECTED] writes:
On Mar 21, 2007, at 5:11 AM, Tom Lane wrote:
constraint_exclusion
Hrm... wasn't that option added in case there was a bug in the
exclusion code?
Well, the bug was a lack of ways to get rid of plans that were
On Mar 26, 2007, at 2:01 AM, Galy Lee wrote:
As AUTOVACUUM is having multiple workers now, the semantics of
autovacuum_cost_limit also need to be redefined.
Currently, autovacuum_cost_limit is the accumulated cost that will
cause
one single worker vacuuming process to sleep. It is used to
See Simon's reply... timestamptz math is *not* IMMUTABLE, because
sessions are free to change their timezone at any time. I bet you can
get some invalid results using that function with a clever test case.
On Mar 26, 2007, at 3:48 PM, Weslee Bilodeau wrote:
Weslee Bilodeau wrote:
Mainly
Jim Nasby [EMAIL PROTECTED] writes:
On Mar 25, 2007, at 12:31 PM, Tom Lane wrote:
The other argument was that you might not want the costs of searching
for contradictory constraints if your workload was such that the
search
never or hardly ever succeeds. That still justifies the existence
On Mon, Apr 09, 2007 at 04:07:16PM -0400, Andrew Dunstan wrote:
Some type systems have named array types, some don't. I can live happily
with either. Are array types anonymous in the standard?
Yes, they're anonymous in the standard. That doesn't mean we can't give
them names if we wanted...
Jim Nasby wrote:
See Simon's reply... timestamptz math is *not* IMMUTABLE, because
sessions are free to change their timezone at any time. I bet you can
get some invalid results using that function with a clever test case.
I'm pretty sure it could easily be broken.
But to make it easier for
Heikki Linnakangas [EMAIL PROTECTED] writes:
I've been working on the patch to enhance our group commit behavior. The
patch is a dirty hack at the moment, but I'm settled on the algorithm
I'm going to use and I know the issues involved.
One question that just came to mind is whether Simon's
On Mon, 9 Apr 2007, Tom Lane wrote:
We put in a workaround a long time ago to make it possible to tell the
difference between btree and hash special space, which are also the same
size: there's an unused 16 bits in hash special space that we fill with
a specific value. As of 8.2 this doesn't
Tom Lane wrote:
Heikki Linnakangas [EMAIL PROTECTED] writes:
I've been working on the patch to enhance our group commit behavior. The
patch is a dirty hack at the moment, but I'm settled on the algorithm
I'm going to use and I know the issues involved.
One question that just came to
Gavin Sherry [EMAIL PROTECTED] writes:
On Mon, 9 Apr 2007, Tom Lane wrote:
... I don't see any way to make it completely bulletproof
without enlarging the special space, which seems an unreasonable price
to pay. But even one chance in 16K is way better than the current
situation.
Sounds
On Mon, 9 Apr 2007, Bruce Momjian wrote:
The big question is who is going to care about the milliseconds delay
and is using a configuration that is going to benefit from commit_delay.
I care. WAL writes are a major bottleneck when many clients are
committing near the same time. Both times
Greg Smith [EMAIL PROTECTED] writes:
An alternate mechanism that tells the client the commit is done when it
hasn't hit disk is of no use for the applications I work with, so I
haven't even been paying attention to no-commit-wait.
Agreed, if you need committed to mean committed then no-wait
On Mon, 9 Apr 2007, Bruce Momjian wrote:
The big question is who is going to care about the milliseconds delay
and is using a configuration that is going to benefit from commit_delay.
I care. WAL writes are a major bottleneck when many clients are
committing near the same time. Both
35 matches
Mail list logo
