a
sequence won't "roll back" in some really bad case, if you fail over
to the target? Figuring out how to do that was one of Jan's homework
projects, IIRC. ;-)
A
--
Andrew Sullivan | [EMAIL PROTECTED]
I remember when computers were frustrating because they *did* exactly wha
e used to
decide what would satisfy our requirements, you can see Jan's concept
paper; it's at
<http://developer.postgresql.org/~wieck/slony1/Slony-I-concept.pdf>.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
This work was visionary and imaginative, and goes to show that visionary
and im
method used by pg_dump so the only throttle will be the network.
Not quite, because your schema needs to be complete on the target
system (in particular, you need your unique keys to stay, although
you can get rid of some other indexes to speed things up).
A
--
Andrew Sullivan | [EMAIL PROTECTED]
; just that
it's important to realise that additional features come at the cost
of additional complexity. My bet is that anyone really needing high
availability will end up needing some of those additional features.)
A
--
Andrew Sullivan | [EMAIL PROTECTED]
When my information changes, I
velopment. We
use it for the .info and .org top level domains, among other systems.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
When my information changes, I alter my conclusions. What do you do sir?
--attr. John Maynard Keynes
---(end of
lem, however, and I can think of a nifty way to attempt this with
the currently-beta Slony-I software.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
The whole tendency of modern prose is away from concreteness.
--George Orwell
---(end of broadcast)
.
If you connect to the local IP (i.e. not 127.0.0.1 but some other
interface), does the same thing happen? (This would tell you whether
the problem lies in some sort of special problem routing localhost,
or whether it's something else.)
A
--
Andrew Sullivan | [EMAIL PROTECTED]
The fact t
ime, the
locks are maintained.
In any case, the answer probably is still to kill -2 the offending
back end.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
The fact that technology doesn't work is no bar to success in the marketplace.
--Philip Greenspun
--
pping support in Slony 1 (HEAD right
now) or the approach outlined at OSCON last year (anonymous file
shipping based on dbmirror -- mighty cool) is something closeish to
what you'll need.
--
Andrew Sullivan | [EMAIL PROTECTED]
This work was visionary and imaginative, and goes to show that vi
l lists.)
Actually, I don't think this _is_ a JDBC issue. AIUI, The problem
bites you with Slony because its triggers use SPI. So the reference
to the oid is to the trigger's oid, which you won't lose until the
back ends are all recyled.
A
--
Andrew Sullivan | [EMAIL PROTECT
x27;re going through the pain of upgrading, might as
well go all the way.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
I remember when computers were frustrating because they *did* exactly what
you told them to. That actually seems sort of quaint now.
--J.D. Baldwin
--
otator, though, if you log all your queries, because your
logs get real big real fast.
That isn't to say that moving off 7.1 isn't a really good idea.
--
Andrew Sullivan | [EMAIL PROTECTED]
The plural of anecdote is not data.
--Roger Brinner
InnoDB. You actually need
to understand the difference between MySQL and its storage engines to
understand why this is the case, and one can't really expect people
who are evaluating different systems to know all that sort of detail
in advance.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
In
re generally, this question has been bandied about here several
times over the past year. Have a look in the archives.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
The fact that technology doesn't work is no bar to success in the marketplace.
--Philip Greenspun
-
not controlled by the InnoDB table handler, whereas we don't
have that problem with MVCC. So under MVCC, by definition, you can't
have partial transaction failures. (Or, more precisely, any such
partial failure is a bug in PostgreSQL, but in MySQL it might be a
feature.)
A
--
Andrew
ts that he's aiming to do this
for 1.1, and I have reason to believe that that'll be well on the
road to release by the end of the year. Chris, are those bits
ready? Maybe Oliver could use them as inspiration for the Debian
work?
A
--
Andrew Sullivan | [EMAIL PROTECTED]
The fact that t
On Fri, Oct 22, 2004 at 09:36:10AM -0400, Jan Wieck wrote:
> wealth. And as long as Afilias is using Slony-I in production, Andrew
> Sullivan will not let me do whatever I want if there's a severe problem
> nobody else can fix.
Or someone around here will, anyway. I might get hi
(not very humble) opinion. I did an
analysis of it this year for my employers. I'm not at liberty to
publish the resulting memo, but I would say that, at the very least,
it's important to understand how the table types, concurrency, and
transaction control all work together. (I beli
om you ask, "Postgres" is either a short form of
PostgreSQL, an ancestor of PostgreSQL, or both.
> Do you know of any places where this information can be obtained?
Here :)
A
--
Andrew Sullivan | [EMAIL PROTECTED]
I remember when computers were frustrating because they *did* exactl
But mapping data points to
places in space is one of those cases where you probably _do_ need
this sort of preallocation mechanism. It's what hotels do, after
all.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
The fact that technology doesn't work is no bar to success i
here you _do_ have gaps.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
In the future this spectacle of the middle classes shocking the avant-
garde will probably become the textbook definition of Postmodernism.
--Brad Holland
---(end of broadcast)-
ver, that you're creating a new problem for yourself by not being
able to skip sequence values. My bet is that you actually need to
find a better way to solve the "other serious problems" you have
rather than banging on sequences to get them to fit your intended
use.
A
--
Andrew Sulliv
u have to throw away the work you did (or present the
intermediate changes to the user, or whatever).
A
--
Andrew Sullivan | [EMAIL PROTECTED]
In the future this spectacle of the middle classes shocking the avant-
garde will probably become the textbook definition of Postmodernism.
be wanting to use slony (www.slony.info).
If you're asking whether Postgres has something akin to Oracle's RAC,
the answer is "not yet".
A
--
Andrew Sullivan | [EMAIL PROTECTED]
This work was visionary and imaginative, and goes to show that visionary
and imaginative work need not end
ySQL appears to beat us on both fronts. Whether such a system is
useful to you is another matter.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
This work was visionary and imaginative, and goes to show that visionary
and imaginative work need not end up well.
--Dennis Ritchie
--
master startup parameters). Isn't there
> any other solution to see what postgresql instances are doing?
/usr/ucb/ps -auxww will work. I've used it thousands of times.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
The fact that technology doesn't work is no bar to succe
27;t
use Slony's own admin tools.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
The fact that technology doesn't work is no bar to success in the marketplace.
--Philip Greenspun
---(end of broadcast)---
TIP 3: if post
s for the
> acl column (e.g. pg_class.relacl) and using that in a function.
Andrew Hammond is about to (has?) post some helper code he has for
managing ACLs more easily.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
I remember when computers were frustrating because they *did* exactly what
you told
On Wed, Aug 11, 2004 at 12:02:07PM -0400, Tom Lane wrote:
> Now erServer did work for them, but it required significant amounts of
> tuning and constant babysitting by the DBA. (If Andrew Sullivan is
> paying attention to this thread, he can offer lots of gory details.)
> I can als
I'll try this again, since it doesn't seem to have made it to the
list.
On Wed, Aug 11, 2004 at 12:02:07PM -0400, Tom Lane wrote:
> Now erServer did work for them, but it required significant amounts of
> tuning and constant babysitting by the DBA. (If Andrew Sullivan is
>
#x27;re using an OS which allows you to set a nocache
option for a filesystem (Solaris allows this on UFS, for instance).
But the kernel cache, in my experience, is _still_ the winner,
because the LRU database cache isn't real efficient at large sizes.
--
Andrew Sullivan
ecause it makes promoting a slave node somewhat risky.
Slony-I has a trick to solve this problem, BTW.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
---(end of broadcast)---
TIP 4: Don't 'kill -9' the postmaster
ugh -- I didn't report it because it wasn't a system I
could leave in that state for debugging.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
---(end of broadcast)---
TIP 8: explain analyze is your friend
ll be ready in the next six months, you might want to check out
the Slony-I project on gborg.postgresql.org. But even that is not
"real time", in that there is a gap between when a tuple is COMMITTED
on the primary node and when it appears on other nodes.
A
--
Andrew Sullivan
starts on the first day of the
month.)
In April 2004, 1 April is Thurs, so
1-7 -> W1
8-14 -> W2
15-21 -> W3
22-28 -> W4
29-30 -> W5 == W1 of May
This is also why 8 May is in week 2 of May, but 7 May is on week 1.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
--
OMMIT;BEGIN; all at once. You get one transaction
through, but you still end up idle in transaciton.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
---(end of broadcast)---
TIP 2: you can get off all lists at once with the unregister command
(send
stake.
So, Oracle Corp offers two different ways to keeo you up nights. :)
I'm sure they're both wonderful products. But they certainly don't
have a one-size-fits-all approach.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
---(end of broadcast)---
TIP 7: don't forget to increase your free space map settings
r jdk and ant? Are they first in your $PATH? I
know that on debian, for instance, when you type 'java' you get
something that's free but which doesn't always work.
--
Andrew Sullivan | [EMAIL PROTECTED]
---(end of broadcast)
On Tue, Apr 20, 2004 at 10:45:33AM -0700, Jeff Boes wrote:
> variable `header'.
> [javac] JTableHeader header = null;
>
> To me, this indicates that the SDK isn't installed (properly). But I
> admit I'm pretty much a Java know-nothin'.
java
get a whole
lot more expensive (7.1's VACUUM is equivalent to later VACUUM
FULL). Why do you have to do this?
A
--
Andrew Sullivan | [EMAIL PROTECTED]
In the future this spectacle of the middle classes shocking the avant-
garde will probably become the textbook definition of Postmoderni
d it imposes a noticable cost to
transactions on the master database.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
---(end of broadcast)---
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to [EMAIL PROTECTED])
On Fri, Mar 26, 2004 at 02:28:10AM -0800, Andrew Mayo wrote:
> PLEASE can we have the native Win32 port SOON.
I don't think anyone is dragging their heels on purpose. If it's
that important to you, consider contributing either in code or in
paid developer time.
A
--
An
7.3, but it likely works for later versions. I have never
used Postgres on Windows (I can barely stand to use Word on Windows),
so I can't help you.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
---(end of broadcast)---
TIP 6: Have you
hould have a look at Chapter 16 of the
docs (see e.g.
<http://www.postgresql.org/docs/7.4/interactive/runtime.html>) for
more detail. Solaris has just unbelievably low defaults for this, so
you'll for sure need to do some work there on every client's machine.
A
--
Andrew Sulliv
you the best
database habits, so I'd urge you to consider the additional work for
Postgres, just because you'll get a good grounding in fundamentals
that way. Postgres is the most rigid of the systems, in that it
usually has a smaller number of well-defined ways to do something.
--
Andrew Su
it is really hard to find out unless brute force is used.)
I have in fact seen IDE drives lie about this, although not recently.
You can be sure it's lying if, even when you turn caching off,
performance never changes.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
---
if
it's set too high and you have enough cases of this, you might
actually cause your box to start swapping.
> and disk speed. I've considered renicing the processes, I was wondering
That is unlikely to help, and certainly won't if the queries are
actually blocked.
--
Andrew
onsidering Postgres to be broken. Am I
> wrong?
Yes, you are. These are different pieces of software, and they'll
have different behaviour. It's not like developers renumber the
versions for fun.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
The fact that technology doesn
o find it)
and then issue kill -2
A
--
Andrew Sullivan | [EMAIL PROTECTED]
This work was visionary and imaginative, and goes to show that visionary
and imaginative work need not end up well.
--Dennis Ritchie
---(end of broadcast)-
Are others having problems with gborg this afternoon?
A
--
Andrew Sullivan
---(end of broadcast)---
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match
ave a design for this sort of thing there,
although it does some remarkably tricky things that not everyone
thinks will work. (I happen to be among the somewhat optimistic on
this front, but I haven't been able to bring any money to the
project.)
A
--
Andrew Sullivan
--
ver on gborg and none of the rest, so I cannot
comment on their relative merits.
A
--
Andrew Sullivan
---(end of broadcast)---
TIP 5: Have you checked our extensive FAQ?
http://www.postgresql.org/docs/faqs/FAQ.html
On Tue, Feb 17, 2004 at 03:39:25PM -0700, Linh Luong wrote:
> Hi All,
>
> 1. I have been reading and the max size of a tuple is 8K.
That hasn't been true for ages. Use a newer PostgreSQL; any modern
one will have the TOAST capability and won't have this limitation.
A
ing. Of course, that's a pretty big
unless, and not one that I'm volunteering to make go away!
A
--
Andrew Sullivan
---(end of broadcast)---
TIP 8: explain analyze is your friend
eron with 40 GB/sec
> bandwidth)
>
> The 32 bit machines cannot compete in these arenas.
I'm not supporting immense databases with this, but I am using 8- and
10- way UltraSPARC II boxes with 16 G of ram. I've been unable to
show a difference. There might _be_ one, mind, I jus
un experts say you should get your
database applications optimised for 64 bits because you can ship
around more data at a time. I have no clue what they're basing it
on, though.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
The plural of anecdote is not data.
I think that your script is mighty dangerous. It sounds
like a recipe for data loss to me. Postgres is considerably more
robust than this, and I think you're trying to cover up some serious
problems that you may have, likely with your hardware.
A
--
Andrew Sullivan
-
rom there. Of course, you can't look at the new code
first, but if Compaq could clean room the IBM BIOS, it's gotta be
possible to find someone who knows nothing about this either.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
In the future this spectacle of the middle classes shocking the av
YourSQL or whatever it was called).
A
--
Andrew Sullivan | [EMAIL PROTECTED]
The fact that technology doesn't work is no bar to success in the marketplace.
--Philip Greenspun
---(end of broadcast)---
TIP 7: don't forget to increase your free space map settings
er reason than that
license wars in general are not really welcome here.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
The plural of anecdote is not data.
--Roger Brinner
---(end of broadcast)---
TIP 4: Don't 'kill -9' the postmaster
e patches.
A
--
Andrew Sullivan
---(end of broadcast)---
TIP 4: Don't 'kill -9' the postmaster
lying about whether the bits are actually on the
disk, you might lose some things you think are committed. You can
apparently tolerate some data loss anyway, so in this case it's not
too big a deal.
A
--
Andrew Sullivan | [EMAIL PROTECTED]
Music is no business of mine.
version?
There's a commercial version as well. It shares some of, but not all
of, the code.
A
--
Andrew Sullivan
It never occurred to them that, if everyone had to think outside the box,
maybe it was the box that needed fixing.
--Malcom Gladwell
ost
you in hardware.
A
--
Andrew Sullivan 204-4141 Yonge Street
Afilias CanadaToronto, Ontario Canada
<[EMAIL PROTECTED]> M2P 2A8
+1 416 646 3304 x110
---
oposal states, the actual SQL is
to be sent in order to the slave. That is, only consistent sets are
sent: you can't have a condition on the slave that never could have
obtained on the master. This means greater overhead for cases where
the same row is altered repeatedly
ll need
oids, which is not standard any more.
If you do your work in one transaction and get the currval that way,
it is impossible to go wrong. Also, if you don't return the
connection to the pool before getting the currval, you will not go
wrong.
A
--
Andrew Sullivan
In fact, !(NULL=NULL) & !(NULL!=NULL).
SQL uses three-valued logic.
A
--
Andrew Sullivan 204-4141 Yonge Street
Afilias CanadaToronto, Ontario Canada
<[EMAIL PROTECTED]> M2P 2A8
The use of this convention is
> extensive.
You could probably put in a rewrite rule to convert '' to NULL and
allow nulls on the column. It's the only suggestion I can think of,
short of going back to 7.2.
A
--
Andrew Sullivan 204-4141 Yo
ariable (and there's the
usual UNIX-y ways of handling it if $PAGER is not set). So you can
make it anything you want.
In 7.3.x and later, you can also do \pset pager to turn off the pager
in psql.
A
--
Andrew Sullivan 204-4141 Yonge Street
Afilias Canada
Postgres logs.
A
--
----
Andrew Sullivan 204-4141 Yonge Street
Afilias CanadaToronto, Ontario Canada
<[EMAIL PROTECTED]> M2P 2A8
+1 416 646 3304 x110
-
od you write it "xType" or xType? With the
quotes, you enforce cases-sensitivity.
> Doesn't it violate documentation:
No. The same page says that "" will enforce the case you wrote.
A
--
Andrew Sullivan 204-4141 Yonge Street
Afilias Cana
he distance that IK
travels is limited :-/
A
--
----
Andrew Sullivan 204-4141 Yonge Street
Afilias CanadaToronto, Ontario Canada
<[EMAIL PROTECTED]> M2P 2A8
+1
n unless
it is "data merge" sort of stuff (e.g. table1 and table 2 on db1;
table 3 and table 4 on db2; and table1, table 2, table3, and table4
on db3).
A
--
Andrew Sullivan 204-4141 Yonge Street
Afilias CanadaToronto, On
3 systems. What are people's feelings
about that? Should we just accept that, as 7.2 is now officially
deprecated, we need not support it?
I'm cc:ing this to the pgsql-general list, because I think several
people who are trying out the code are not on this list.
A
--
Andrew Sulliv
this topic in the archives.
If you need an approximate value, you can get it from the system
tables.
A
--
----
Andrew Sullivan 204-4141 Yonge Street
Afilias CanadaToronto, Ontario Canada
<[EMAIL PROTECTED]>
the "growing pains" argument instead? What are
those arguments, anyway? ( I think I know, but maybe not.)
A
--
Andrew Sullivan 204-4141 Yonge Street
Afilias CanadaToronto, Ontario Canada
<[EMAIL PROTECTED]>
anage multiple
installations and back ends in our test and production environments.
It works pretty well. I also know that Tom Lane uses something again
similar in order to support the apparently limitless different
permutations of installations he has available at the touch of '.
filename
odules. Ask 'em about the reliability
of replacement parts.
A
--
Andrew Sullivan 204-4141 Yonge Street
Liberty RMS Toronto, Ontario Canada
<[EMAIL PROTECTED]> M2P 2A8
r
barrier for corporate use. One can only do so much testing, and it's
always possible you've missed something. You need to be able to go
back to some known-working state.
A
--
Andrew Sullivan 204-4141 Yonge Street
Liberty RMS
ince they post all
the contracts.
A
----
Andrew Sullivan 204-4141 Yonge Street
Liberty RMS Toronto, Ontario Canada
<[EMAIL PROTECTED]> M2P 2A8
+1 41
you need
to.
There's an erserver mailing list, BTW. I'm on it.
A
--
Andrew Sullivan 204-4141 Yonge Street
Liberty RMS Toronto, Ontario Canada
<[EMAIL PROTECTED]> M2P 2A8
r as well. All (?) are
found on gborg. I can't speak about the others, but erserver is a
little bit painful to set up. Works, though.
A
--
Andrew Sullivan 204-4141 Yonge Street
Liberty RMS Toronto, Ontario Cana
compiler installed, (b) your compiler isn't in your $PATH, or (c)
some necessary libraries are not available for your compiler to run.
But without seeing the real error message, I can't help you.
A
--
Andrew Sullivan 204-4141 Yonge Street
Liberty RMS
r the opportunity.
Well, since the main point was to get some $$ into the company, bucks
which IBM has and PostgreSQL doesn't, it's not too surprising that
the PostgreSQL team didn't win. The move to DB2 was apparently a
quid pro quo for the cash.
A
--
Andrew Sullivan
led vacuum isn't it?
Not for some cases. In (say) 40% write situation, you have _lots_ of
dead tuples. Perhaps you can make the application more efficient,
but that's not always an option (maybe you don't have the code).
A
--
Andrew Sullivan
s
> waiting on it.
Right, but all that intelligence is something that isn't in there
now. And anyway, the real issue is I/O, not processor.
A
--
Andrew Sullivan 204-4141 Yonge Street
Liberty RMS Toronto, Ontario Ca
een our test env. and the
production env., so we can put in a fudge factor for this; but that's
it.
A
--
Andrew Sullivan 204-4141 Yonge Street
Liberty RMS Toronto, Ontario Canada
<[EMAIL PROTECTED]>
On Wed, Aug 20, 2003 at 05:30:14PM -0400, Joseph Tate wrote:
> Is there a way to get a list of all the currently open transactions on a
> database?
You can use the statistics in pg_stat_activity if you have the query
stats turned on.
A
--
Andrew Sullivan 20
or time -- in which
case everybody else gets their processor slice but can't do anything,
because they have to wait until the niced vacuum process gets back in
line.
A
--
----
Andrew Sullivan 204-4141 Yonge Street
Liberty RMS Toronto, Onta
do
> the same thing.
Assuming that one can keep up with the dust bunnies this way, though,
one wouldn't need to do vacuum full. This would definitely be a way
cool feature, if implementable.
A
--
Andrew Sullivan 204-4141 Yonge Street
Liberty RMS
for Sun (and conversely) partly
because, I expect, that's where a lot of their users were, and the
performance or reliability gains were significant. Whether that is
worth doing for PostgreSQL, when there are probably lots of other
targets to aim at, is an open question.
A
--
Andrew
ough -- it seems to me what would be more useful is an even
lazier vacuum: something that could be told "clean up as cycles are
available, but make sure you stay out of the way." Of course, that's
easy to say glibly, and mighty hard to do, I expect.
A
--
info/asullivan.pdf>.
A
--
Andrew Sullivan 204-4141 Yonge Street
Liberty RMS Toronto, Ontario Canada
<[EMAIL PROTECTED]> M2P 2A8
+1 416 646 3304 x110
not related.
> And is rserv limited to a demo or is it a full solution ?
contrib/rserv, just like all the other code in the tarball, is free
software.
A version of replication that descended from rserv is going to appear
on gborg someday soon. I'm not sure when
r?
I think you need the .pgpass in the home directory of the user
running the server processing the PHP (e.g. ~apache or ~nobody or
whatever you use). I'm not sure how well it will work, though, given
that those users typically have /bin/false as a shell. I haven't
tested it.
less all the databases are shut down. You could use pg_dump.
A
--
Andrew Sullivan 204-4141 Yonge Street
Liberty RMS Toronto, Ontario Canada
<[EMAIL PROTECTED]> M2P 2A8
I think your administrator is talking out of his hat. We switch back
and forth between Linux and Solaris all the time here, and PostgreSQL
works the same all the time.
A
--
Andrew Sullivan 204-4141 Yonge Street
Liberty RMS Toronto, Ontario Ca
ame connection.
> According to our postgres adminstrator, Postgres seems
> to behave differently on Linux and Solaris. Any ideas on that?
Differently how? Start up time for a connection is sure worse:
fork() on Solaris is slow as my dog with his foot in a bandage.
A
--
Andrew Sull
l upper case. Postgres will
automatically fold them to lower case anyway, but you won't get them
displayed in all upper case.
A
--
Andrew Sullivan 204-4141 Yonge Street
Liberty RMS Toronto, Ontario Canada
<[EMAIL PROTECTED]&g
t have a db might want to export his data through his web
> interface.
Hmm. Well, you could use a super-user account to do the dump. It
would have access to everything, and would only need one password.
I'm wondering about security implications of that, however.
A
--
Andrew Sulliva
601 - 700 of 733 matches
Mail list logo