Heikki Linnakangas heikki.linnakan...@enterprisedb.com writes:
Tom Lane wrote:
There's no equivalent of XLogArchivingActive()?
XLogArchivingMode() == false enables us to skip WAL-logging in
operations like CLUSTER or COPY, which is a big optimization. I don't
see anything like that in Hot
Tom Lane wrote:
Heikki Linnakangas heikki.linnakan...@enterprisedb.com writes:
Tom Lane wrote:
There's no equivalent of XLogArchivingActive()?
XLogArchivingMode() == false enables us to skip WAL-logging in
operations like CLUSTER or COPY, which is a big optimization. I don't
see anything
On Sat, 2009-11-14 at 08:43 -0800, Robert Hodges wrote:
I can help set up automated basic tests for hot standby using 1+1 setups on
Amazon. I¹m already working on tests for warm standby for our commercial
Tungsten implementation and need to solve the problem of creating tests that
adapt
Simon Riggs wrote:
On Sat, 2009-11-14 at 14:59 +0200, Heikki Linnakangas wrote:
I can't see any obvious way around that.
Huh? We're only doing this strict locking approach because you insisted
that the looser approach was not acceptable.
Take it easy, Simon. By obvious, I meant trivial or
Oh, forgot to mention another thing that I've been pondering:
Currently, the running-xacts record is written to the WAL after the
checkpoint record. There's a small chance that you get an xlog switch in
between. If that happens, it might take a long time after the checkpoint
record until the
Simon Riggs wrote:
On Sun, 2009-11-15 at 21:37 +0200, Heikki Linnakangas wrote:
Am I missing anything?
Will review.
Thanks! Please use the head of git branch, I already found one major
oversight in what I posted that's fixed there... I should go to bed already.
--
Heikki Linnakangas
On 11/15/09 2:25 AM PST, Simon Riggs si...@2ndquadrant.com wrote:
On Sat, 2009-11-14 at 08:43 -0800, Robert Hodges wrote:
I can help set up automated basic tests for hot standby using 1+1 setups on
Amazon. I¹m already working on tests for warm standby for our commercial
Tungsten
Hi Simon and Heikki,
I can help set up automated basic tests for hot standby using 1+1 setups on
Amazon. I¹m already working on tests for warm standby for our commercial
Tungsten implementation and need to solve the problem of creating tests that
adapt flexibly across different replication
On Fri, 2009-11-13 at 22:19 +0200, Heikki Linnakangas wrote:
I got the impression earlier that you had some test environment set up
to test hot standby. Can you share any details of what test cases
you've run?
Fair question. The Sep 15 submission happened too quickly for us to
mobilise
Devrim GÜNDÜZ wrote:
Is there a reason why recovery.conf.sample does not include (sample)
entries for recovery_connections and max_standby_delay?
No, they probably should be included. I'll add them, thanks.
--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com
--
Sent via
On Tue, 2009-11-10 at 22:00 +0200, Devrim GÜNDÜZ wrote:
On Tue, 2009-11-10 at 20:36 +0200, Heikki Linnakangas wrote:
Attached is the latest and greatest patch against CVS head, taken from
the hs-riggs branch in my git repository.
Is there a reason why recovery.conf.sample does not include
On Tue, 2009-11-10 at 15:11 -0500, Michael Glaesemann wrote:
I skimmed through the documentation to get a better handle on what
this will mean.
Thanks for this and any further corrections/additions.
--
Simon Riggs www.2ndQuadrant.com
--
Sent via pgsql-hackers mailing list
On Wed, 2009-11-11 at 12:35 +0200, Heikki Linnakangas wrote:
While reading through the patch for what must be the 100th time by now,
:-)
it occurred to me that this comment in heap_xlog_freeze:
+ /*
+* Freezing tuples does not require conflict processing
+*/
is
On Tue, 2009-11-10 at 20:36 +0200, Heikki Linnakangas wrote:
Attached is the latest and greatest patch against CVS head, taken from
the hs-riggs branch in my git repository.
Is there a reason why recovery.conf.sample does not include (sample)
entries for recovery_connections and
On Nov 10, 2009, at 13:36 , Heikki Linnakangas wrote:
Attached is the latest and greatest patch against CVS head, taken from
the hs-riggs branch in my git repository.
Awesome. Thank you, Simon and Heikki!
I skimmed through the documentation to get a better handle on what
this will mean.
Michael Glaesemann wrote:
Other commands are in all-caps. Any reason INSERT, UPDATE, DELETE,
TRUNCATE, CREATE, DROP, ALTER, and COMMENT are not?
No. Thanks, I'll fix that.
--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com
--
Sent via pgsql-hackers mailing list
Simon Riggs wrote:
On Thu, 2009-10-08 at 11:50 -0400, Tom Lane wrote:
Robert Haas robertmh...@gmail.com writes:
Subcommitting every single row is going to be really painful,
especially after Hot Standby goes in and we have to issue a WAL record
after every 64 subtransactions (AIUI).
Andrew Dunstan and...@dunslane.net writes:
Simon Riggs wrote:
HS writes a WAL record for subtransactions at the point that the subxid
cache overflows for any single transaction. Current cache size = 64.
Top-level transaction then writes one additional WAL record every
additional 64 subxids
On Thu, 2009-10-22 at 07:55 +0300, Heikki Linnakangas wrote:
Simon Riggs wrote:
On Wed, 2009-10-21 at 23:02 +0300, Heikki Linnakangas wrote:
Hmm, dunno about that, but there is one problem with the grant to dummy
proc, then release in startup process approach. What if there isn't
enough
Simon Riggs wrote:
On Thu, 2009-10-22 at 07:55 +0300, Heikki Linnakangas wrote:
Making some effort to transfer locks instead of acquiring+releasing
would eliminate the need for having extra lock space available when
switching from hot standby mode to normal operation.
This isn't very clear.
On Thu, 2009-10-22 at 09:41 +0300, Heikki Linnakangas wrote:
Simon Riggs wrote:
On Thu, 2009-10-22 at 07:55 +0300, Heikki Linnakangas wrote:
Making some effort to transfer locks instead of acquiring+releasing
would eliminate the need for having extra lock space available when
switching
On Wed, 2009-10-21 at 19:37 +0300, Heikki Linnakangas wrote:
So, I'm quite eager to just revert all those lock_twophase_recover()
changes, and always rely on the grant lock to dummy proc, then
release
it in startup process method. If we don't want to rely on that,
PostPrepare_Locks is an
Simon Riggs wrote:
On Wed, 2009-10-21 at 19:37 +0300, Heikki Linnakangas wrote:
So, I'm quite eager to just revert all those lock_twophase_recover()
changes, and always rely on the grant lock to dummy proc, then
release
it in startup process method. If we don't want to rely on that,
On Wed, 2009-10-21 at 23:02 +0300, Heikki Linnakangas wrote:
Simon Riggs wrote:
On Wed, 2009-10-21 at 19:37 +0300, Heikki Linnakangas wrote:
So, I'm quite eager to just revert all those lock_twophase_recover()
changes, and always rely on the grant lock to dummy proc, then
release
it
Simon Riggs wrote:
On Wed, 2009-10-21 at 23:02 +0300, Heikki Linnakangas wrote:
Hmm, dunno about that, but there is one problem with the grant to dummy
proc, then release in startup process approach. What if there isn't
enough shared memory available to re-acquire the lock for the dummy
proc?
Simon Riggs wrote:
On Fri, 2009-10-16 at 09:33 +0300, Heikki Linnakangas wrote:
If you pause recovery, and then continue, we reset to target none
mode, even if a stopping point was set previously.
Yes, currently. Resetting the target mode will do what you want, rather
than continue.
On Tue, 2009-10-20 at 10:33 +0300, Heikki Linnakangas wrote:
At this point, I'd like to cut out all those control functions to
pause/stop at various points from the patch.
OK
--
Simon Riggs www.2ndQuadrant.com
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
On Tue, 2009-10-20 at 09:43 +0100, Simon Riggs wrote:
On Tue, 2009-10-20 at 10:33 +0300, Heikki Linnakangas wrote:
At this point, I'd like to cut out all those control functions to
pause/stop at various points from the patch.
OK
Maybe OK, I should say. Some parts are important for testing
On Fri, 2009-10-16 at 09:33 +0300, Heikki Linnakangas wrote:
In advance-mode, we will merrilly skip over a WAL record that's a
recovery stop target. Is that a bug or a feature?
Merrily?!? I saw it more as a sombre stepping motion.
Advance currently means set target at next record and
On Thu, 2009-10-15 at 10:33 +0300, Heikki Linnakangas wrote:
There's been a lot of churn in hot standby since the beginning of the
commitfest, so I thought it would be good to summarize where we are.
Attached is the latest and greatest patch against CVS head, taken from
the hs-riggs branch
Simon Riggs wrote:
On Thu, 2009-10-15 at 10:33 +0300, Heikki Linnakangas wrote:
Let me know if I'm missing something. And please feel free to help, by
testing, by reviewing and commenting on the patch, or by addressing any
of the above issues. I will continue working on this, but this is a big
On Fri, 2009-10-16 at 01:43 +0300, Heikki Linnakangas wrote:
recovery_starts_paused is useless as it is. It pauses the recovery right
after the first WAL record, all right, but before we see a running-xacts
record, we won't let any backends in. And if you can't connect, you
can't unpause, so
On Wed, 2009-10-14 at 14:43 +0300, Heikki Linnakangas wrote:
In the hot standby patch, we have this comment in procarray.c:
It is
* important that the XLOG_XACT_ASSIGNMENT record contain *all* subxids
* not just those so far unreported because the sole purpose is to ensure
* we can
Simon Riggs wrote:
You have posted patches that I have said I don't agree with. My name is
going to be on this when it goes in, so I don't think it makes any sense
to force that commit to include changes I don't agree with. I cannot
prevent you making changes afterwards, nor would I wish to.
While playing with conflict resolution, I bumped into this:
postgres=# begin ISOLATION LEVEL SERIALIZABLE;
BEGIN
postgres=# SELECT * FROM foo;
id | data
+--
12 |
(1 row)
postgres=# SELECT * FROM foo;
id | data
+--
12 |
(1 row)
postgres=# SELECT * FROM foo;
id | data
Simon Riggs wrote:
Tom Lane wrote:
[ scratches head ... ] Why is hot standby messing with this sort of
thing at all? It sounds like a performance optimization that should
be considered separately, and *later*.
Yeah, I too considered just ripping it out. Simon is worried that
locking all
On Oct 9, 2009, at 1:21 PM, Bruce Momjian br...@momjian.us wrote:
Simon Riggs wrote:
You have posted patches that I have said I don't agree with. My
name is
going to be on this when it goes in, so I don't think it makes any
sense
to force that commit to include changes I don't agree with.
Robert Haas wrote:
But at least for simple features I think that there would be a value
in separating the patch author's work from the committer's adjustments.
That is just going to make life harder for committers.
There are plenty of things with my name on them that are not exactly
On Fri, 2009-10-09 at 14:05 -0400, Andrew Dunstan wrote:
Robert Haas wrote:
But at least for simple features I think that there would be a value
in separating the patch author's work from the committer's adjustments.
That is just going to make life harder for committers.
There
Andrew Dunstan wrote:
Robert Haas wrote:
But at least for simple features I think that there would be a value
in separating the patch author's work from the committer's adjustments.
That is just going to make life harder for committers.
There are plenty of things with my name
On Wed, 2009-10-07 at 23:19 -0400, Heikki Linnakangas wrote:
Heikki Linnakangas wrote:
I've made public the version I'm working on. That's the version I'm
ultimately going to commit. It would be a lot more helpful if you
provided these patches over that version. Otherwise I have to
Simon Riggs wrote:
I would still like you to make a clear statement that the contents of
that repository are BSD licenced open source contributions.
Ok. All the content in the repository at
git://git.postgresql.org/git/users/heikki/postgres.git is released under
the same BSD license as
On Thu, 2009-10-08 at 07:33 -0400, Heikki Linnakangas wrote:
Simon Riggs wrote:
I would still like you to make a clear statement that the contents of
that repository are BSD licenced open source contributions.
Ok. All the content in the repository at
On Thu, 2009-10-08 at 12:49 +0100, Simon Riggs wrote:
I'll get cracking on some changes.
This will probably be next week now, just in case you're wondering when
I'll start adding patches.
--
Simon Riggs www.2ndQuadrant.com
--
Sent via pgsql-hackers mailing list
On Fri, 2009-10-02 at 10:32 +0300, Heikki Linnakangas wrote:
Simon Riggs wrote:
I will docuemnt the recommendation to set max_standby_delay = 0 if
performing an archive recovery (and explain why).
Hmm, not sure if that's such a good piece of advice either. It will mean
waiting for
On Tue, 2009-09-22 at 12:53 +0300, Heikki Linnakangas wrote:
It looks like the standby tries to remove XID 4323 from the
known-assigned hash table, but it's not there because it was removed
and set in pg_subtrans by an XLOG_XACT_ASSIGNMENT record earlier. I
guess we should just not throw an
On Wed, 2009-09-23 at 12:07 +0300, Heikki Linnakangas wrote:
we need be careful to avoid putting any extra work into the normal
recovery path. Otherwise bugs in hot standby related code can cause
crash recovery to fail.
Re-checked code and found a couple of additional places that needed
tests
Simon Riggs wrote:
On Tue, 2009-09-22 at 12:53 +0300, Heikki Linnakangas wrote:
It looks like the standby tries to remove XID 4323 from the
known-assigned hash table, but it's not there because it was removed
and set in pg_subtrans by an XLOG_XACT_ASSIGNMENT record earlier. I
guess we should
Heikki Linnakangas wrote:
I've made public the version I'm working on. That's the version I'm
ultimately going to commit. It would be a lot more helpful if you
provided these patches over that version. Otherwise I have to refactor
them over that codebase, possibly introducing new bugs.
On Mon, 2009-10-05 at 18:30 -0400, Heikki Linnakangas wrote:
Simon Riggs wrote:
On Mon, 2009-09-28 at 11:25 +0300, Heikki Linnakangas wrote:
Heikki Linnakangas wrote:
Per Simon's request, for the benefit of the archive, here's all the
changes I've done on the patch since he posted the
On Tue, 2009-10-06 at 01:10 +0300, Heikki Linnakangas wrote:
Simon Riggs wrote:
We discussed briefly your change
0011-Replace-per-proc-counters-of-loggable-locks-with-per.patch.
I don't see how that helps at all. The objective of lock counters was to
know if we can skip acquiring an
On Sun, 2009-09-27 at 13:57 +0300, Heikki Linnakangas wrote:
Per Simon's request, for the benefit of the archive, here's all the
changes I've done on the patch since he posted the initial version for
review for this commitfest as incremental patches. This is extracted
from my git repository
On Thu, 2009-10-01 at 18:47 +0300, Heikki Linnakangas wrote:
And if you could please review the changes I've been doing, just to
make sure I haven't inadvertently introduced new bugs. That has
happened before, as you've rightfully reminded me :-).
You posted 17 patches here.
I've
On Mon, 2009-09-28 at 11:25 +0300, Heikki Linnakangas wrote:
Heikki Linnakangas wrote:
Per Simon's request, for the benefit of the archive, here's all the
changes I've done on the patch since he posted the initial version for
review for this commitfest as incremental patches. This is
On Sun, 2009-09-27 at 13:57 +0300, Heikki Linnakangas wrote:
Per Simon's request, for the benefit of the archive, here's all the
changes I've done on the patch since he posted the initial version for
review for this commitfest as incremental patches. This is extracted
from my git repository
Simon Riggs si...@2ndquadrant.com writes:
I don't see how that helps at all. The objective of lock counters was to
know if we can skip acquiring an LWlock on all lock partitions. This
change keeps the lock counters yet acquires the locks we were trying to
avoid. This change needs some
On Mon, 2009-10-05 at 10:19 -0400, Tom Lane wrote:
Simon Riggs si...@2ndquadrant.com writes:
I don't see how that helps at all. The objective of lock counters was to
know if we can skip acquiring an LWlock on all lock partitions. This
change keeps the lock counters yet acquires the locks
Simon Riggs wrote:
We discussed briefly your change
0011-Replace-per-proc-counters-of-loggable-locks-with-per.patch.
I don't see how that helps at all. The objective of lock counters was to
know if we can skip acquiring an LWlock on all lock partitions. This
change keeps the lock counters
Simon Riggs wrote:
On Mon, 2009-09-28 at 11:25 +0300, Heikki Linnakangas wrote:
Heikki Linnakangas wrote:
Per Simon's request, for the benefit of the archive, here's all the
changes I've done on the patch since he posted the initial version for
review for this commitfest as incremental
Simon Riggs wrote:
@@ -7061,6 +7061,15 @@ ShutdownXLOG(int code, Datum arg)
else
{
/*
+* Take a snapshot of running transactions and write this to WAL.
+* This allows us to reconstruct the state of running transactions
+* during archive recovery, if
Simon Riggs wrote:
I will docuemnt the recommendation to set max_standby_delay = 0 if
performing an archive recovery (and explain why).
Hmm, not sure if that's such a good piece of advice either. It will mean
waiting for queries forever, which probably isn't what you want if
you're performing
On Fri, 2009-10-02 at 10:04 +0300, Heikki Linnakangas wrote:
Simon Riggs wrote:
@@ -7061,6 +7061,15 @@ ShutdownXLOG(int code, Datum arg)
else
{
/*
+* Take a snapshot of running transactions and write this to WAL.
+* This allows us to reconstruct the
Simon Riggs wrote:
On Fri, 2009-10-02 at 10:04 +0300, Heikki Linnakangas wrote:
Simon Riggs wrote:
@@ -7061,6 +7061,15 @@ ShutdownXLOG(int code, Datum arg)
else
{
/*
+* Take a snapshot of running transactions and write this to WAL.
+* This allows us to
On Fri, 2009-10-02 at 10:32 +0300, Heikki Linnakangas wrote:
Simon Riggs wrote:
I will docuemnt the recommendation to set max_standby_delay = 0 if
performing an archive recovery (and explain why).
Hmm, not sure if that's such a good piece of advice either. It will mean
waiting for
On Fri, 2009-10-02 at 10:43 +0300, Heikki Linnakangas wrote:
It seems dangerous to write a WAL record after the shutdown checkpoint.
It will be overwritten by subsequent startup, which is a recipe for trouble.
I've said its a corner case and not worth spending time on. I'm putting
it in at
Simon Riggs wrote:
On Fri, 2009-10-02 at 10:43 +0300, Heikki Linnakangas wrote:
It seems dangerous to write a WAL record after the shutdown checkpoint.
It will be overwritten by subsequent startup, which is a recipe for trouble.
I've said its a corner case and not worth spending time on.
On Fri, 2009-10-02 at 11:26 +0300, Heikki Linnakangas wrote:
Simon Riggs wrote:
On Fri, 2009-10-02 at 10:43 +0300, Heikki Linnakangas wrote:
It seems dangerous to write a WAL record after the shutdown checkpoint.
It will be overwritten by subsequent startup, which is a recipe for
On Fri, 2009-10-02 at 11:26 +0300, Heikki Linnakangas wrote:
Simon Riggs wrote:
On Fri, 2009-10-02 at 10:43 +0300, Heikki Linnakangas wrote:
It seems dangerous to write a WAL record after the shutdown checkpoint.
It will be overwritten by subsequent startup, which is a recipe for
Simon Riggs wrote:
I'd rather just skip this for now. It's a minor case anyway and there's
nothing stopping writing their own RunningXactData records with a
function, if it is needed. I can add a function for that.
That won't help. There's no way to have it in a right place wrt. the
shutdown
On Fri, 2009-10-02 at 13:52 +0300, Heikki Linnakangas wrote:
Simon Riggs wrote:
I'd rather just skip this for now. It's a minor case anyway and there's
nothing stopping writing their own RunningXactData records with a
function, if it is needed. I can add a function for that.
That won't
Simon Riggs wrote:
I will add code to make a shutdown checkpoint be a valid starting place
for Hot Standby, as long as there are no in-doubt prepared transactions.
That way we know there are no xids still running and no locks, without
needing to write a record to say so.
Ok, I can live with
On Sun, 2009-09-27 at 14:59 +0300, Heikki Linnakangas wrote:
The locking in smgr_redo_commit and smgr_redo_abort doesn't look right.
First of all, smgr_redo_abort is not holding XidGenLock and
ProcArrayLock while modifying ShmemVariableCache-nextXid and
ShmemVariableCache-latestCompletedXid,
On Fri, 2009-10-02 at 18:04 +0300, Heikki Linnakangas wrote:
Simon Riggs wrote:
I will add code to make a shutdown checkpoint be a valid starting place
for Hot Standby, as long as there are no in-doubt prepared transactions.
That way we know there are no xids still running and no locks,
+ /*
+* If our initial RunningXactData had an overflowed snapshot then we
+* knew we were missing some subxids from our snapshot. We can use
+* this data as an initial snapshot, but we cannot yet mark it valid.
+* We know that the missing subxids are equal to or earlier than
+
On Wed, 2009-09-30 at 18:45 +0300, Heikki Linnakangas wrote:
Regarding this in InitStandbyDelayTimers:
+ /*
+* If replication delay is enormously huge, just treat that as
+* zero and work up from there. This prevents us from acting
+* foolishly when replaying old log files.
+
On Sun, 2009-09-27 at 15:35 +0300, Heikki Linnakangas wrote:
TransactionIdIsInProgress() doesn't consult the known-assigned-xids
structure. That's a problem: in the standby, TransactionIdIsInProgress()
can return false for a transaction that is still running in the master.
HeapTupleSatisfies*
On Thu, 2009-10-01 at 14:29 +0300, Heikki Linnakangas wrote:
+ /*
+* If our initial RunningXactData had an overflowed snapshot then we
+* knew we were missing some subxids from our snapshot. We can use
+* this data as an initial snapshot, but we cannot yet mark it valid.
+*
On Wed, 2009-09-30 at 09:33 +0300, Heikki Linnakangas wrote:
Looking at the changes to StartupMultiXact, you're changing the locking
so that both MultiXactOffsetControlLock and MultiXactMemberControlLock
are acquire first before changing anything. Why? Looking at the other
functions in that
Simon Riggs wrote:
On Wed, 2009-09-30 at 09:33 +0300, Heikki Linnakangas wrote:
Looking at the changes to StartupMultiXact, you're changing the locking
so that both MultiXactOffsetControlLock and MultiXactMemberControlLock
are acquire first before changing anything. Why? Looking at the other
Simon Riggs wrote:
Hmm, yes. ISTM that I'm still calculating latestRunningXid the old way
while assuming it is calculated the new way. The new way is just to grab
nextXid since we have XidGenLock and do TransactionIdRetreat() on it.
Ok, good, that's what I thought too. I'll fix that.
--
On Thu, 2009-10-01 at 18:47 +0300, Heikki Linnakangas wrote:
Could you look into these two TODO items you listed on the wiki page:
Unless we agree otherwise, if its listed on the Wiki page then I will
work on it.
Maybe not as when you might like it, but I am working through the list.
5 new
On Thu, 2009-10-01 at 18:48 +0300, Heikki Linnakangas wrote:
Simon Riggs wrote:
Hmm, yes. ISTM that I'm still calculating latestRunningXid the old way
while assuming it is calculated the new way. The new way is just to grab
nextXid since we have XidGenLock and do TransactionIdRetreat() on
Looking at the changes to StartupMultiXact, you're changing the locking
so that both MultiXactOffsetControlLock and MultiXactMemberControlLock
are acquire first before changing anything. Why? Looking at the other
functions in that file, all others that access both files are happy to
acquire one
Looking at the changes to StartupMultiXact, you're changing the locking
so that both MultiXactOffsetControlLock and MultiXactMemberControlLock
are acquired first before changing anything. Why? Looking at the other
functions in that file, all others that access both files are happy to
acquire one
Regarding this in InitStandbyDelayTimers:
+ /*
+* If replication delay is enormously huge, just treat that as
+* zero and work up from there. This prevents us from acting
+* foolishly when replaying old log files.
+*/
+ if (*currentDelay_ms 0)
+ *currentDelay_ms = 0;
+
Heikki Linnakangas wrote:
Per Simon's request, for the benefit of the archive, here's all the
changes I've done on the patch since he posted the initial version for
review for this commitfest as incremental patches. This is extracted
from my git repository at
Per Simon's request, for the benefit of the archive, here's all the
changes I've done on the patch since he posted the initial version for
review for this commitfest as incremental patches. This is extracted
from my git repository at
git://git.postgresql.org/git/users/heikki/postgres.git.
--
The locking in smgr_redo_commit and smgr_redo_abort doesn't look right.
First of all, smgr_redo_abort is not holding XidGenLock and
ProcArrayLock while modifying ShmemVariableCache-nextXid and
ShmemVariableCache-latestCompletedXid, respectively, like
smgr_redo_commit is. Attached patch fixes that.
TransactionIdIsInProgress() doesn't consult the known-assigned-xids
structure. That's a problem: in the standby, TransactionIdIsInProgress()
can return false for a transaction that is still running in the master.
HeapTupleSatisfies* functions can incorrectly set HEAP_XMIN/XMAX_INVALID
hint bits
On Sat, Sep 26, 2009 at 5:49 AM, Simon Riggs si...@2ndquadrant.com wrote:
Just a note to say that Hot Standby patch is now on git repository
git://git.postgresql.org/git/users/simon/postgres
Branch name: hot_standby
Awesome! Thanks for taking the time to get this set up.
The complete
On Sat, 2009-09-26 at 09:29 -0400, Robert Haas wrote:
I estimate that making the remaining changes noted on the Wiki and
fully
testing them will take at least 2 weeks. Gabriele Bartolini is assisting
in this area, though neither of us are able to work full time on this.
We still have
On 09/26/2009 10:04 AM, Simon Riggs wrote:
If you think there's
something useful I could do, let me know and I'll take a look.
I feel like I need a better way of unit testing new code. Some of the
code in the patch is to handle corner cases, so recreating them is
fairly hard. It is a
On Sat, Sep 26, 2009 at 10:45:17AM -0400, Mark Mielke wrote:
On 09/26/2009 10:04 AM, Simon Riggs wrote:
If you think there's
something useful I could do, let me know and I'll take a look.
I feel like I need a better way of unit testing new code. Some of the
code in the patch is to handle
I feel like I need a better way of unit testing new code. Some of the
code in the patch is to handle corner cases, so recreating them is
fairly hard. It is a nagging feeling that I am missing some knowledge
here and would welcome some insight, or research, into better ways of
doing general
On 09/26/2009 02:28 PM, Dan Colish wrote:
There are a variety of projects dedicated to creating C unit test
frameworks. I don't have a lot of experience with them, but I have heard
good things about check and cunit. Here's a link I found with a longer
list of frameworks.
On Sep 26, 2009, at 12:33 PM, Josh Berkus wrote:
There's always pgtap. Whenever we find a new corner case, we add it
to
the development test suite.
Also, for C TAP, there's [libtap](http://jc.ngo.org.uk/trac-bin/trac.cgi/wiki/LibTap
). You can then use `prove` which you likely already
Mark Mielke escribió:
Most real life code gets a little more complicated. For example,
what if we want to simulate a network failure or out of disk space
condition? What if we want to test out what happens when the Y2038
date is reached? This requires either complex test case setup that
is
Looking at the startup sequence now.
I see that you modified ExtendSUBTRANS so that it doesn't wipe out
previously set values if it's called with out-of-order xids. I guess
that works, although I think it can leave pages unzeroed if it's called
with a large enough gap between xids, so that the
On Thu, 2009-09-24 at 13:33 +0300, Heikki Linnakangas wrote:
Heikki Linnakangas wrote:
The problem becomes a lot easier if we accept that it's OK to have a
lock included in the running-xacts snapshot and also appear in a
XLOG_RELATION_LOCK record later. The standby should handle that
On Wed, 2009-09-23 at 19:07 +0300, Heikki Linnakangas wrote:
Rather than keep the numHeldLocks counters per-proc in proc array, I
think it would be simpler to have a single (or one per lock partition)
counter in shared memory in lock.c. It's just an optimization to make it
faster to find out
401 - 500 of 968 matches
Mail list logo