> the sorting all a relation's early pages are going to be in "in a row".
>
Not sure, what is best way to tackle this problem, but I think one way could
be to perform sorting at flush requests level rather than before writing
to OS buffers.
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
On Tue, Jan 12, 2016 at 5:52 PM, Andres Freund wrote:
>
> On 2016-01-12 17:50:36 +0530, Amit Kapila wrote:
> > On Tue, Jan 12, 2016 at 12:57 AM, Andres Freund
wrote:>
> > >
> > > My theory is that this happens due to the sorting: pgbench is an
update
> > &g
On Tue, Jan 12, 2016 at 7:24 PM, Andres Freund wrote:
> On 2016-01-12 19:17:49 +0530, Amit Kapila wrote:
> > Why can't we do it at larger intervals (relative to total amount of
writes)?
> > To explain, what I have in mind, let us assume that checkpoint interval
> > is l
On Mon, Jan 11, 2016 at 9:16 AM, Amit Kapila
wrote:
> On Mon, Jan 11, 2016 at 3:14 AM, Peter Geoghegan wrote:
> >
> > On Sun, Jan 10, 2016 at 9:13 AM, Robert Haas
wrote:
> > >> I'm not sure why the test for nworkers following the
> > >> LaunchP
compare and swap operation.
>
> After above fix memory corruption is not visible. But I see some more
> failures at higher client sessions(128 it is easily reproducible).
>
>
Don't you think we need to update the snapshot fields like count,
subcount before saving i
to a snapshot, so these xids will not be reported as "running"
by this function. This is OK for current uses, because we always check
TransactionIdIsCurrentTransactionId first, except for known-committed
XIDs which could not be ours anyway.
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
gress
in the server by checking buffers dirtied/written (we can refer that
information using pgBufferUsage) since last time we log this record in
bgwriter.
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
On Sat, Jan 16, 2016 at 5:08 PM, Michael Paquier
wrote:
>
> On Sat, Jan 16, 2016 at 7:10 PM, Amit Kapila
wrote:
> > On Sun, Dec 20, 2015 at 6:44 PM, Michael Paquier <
michael.paqu...@gmail.com>
> > wrote:
> >> On Sun, Nov 8, 2015 at 9:50 PM, Michael Paquier
On Sat, Jan 16, 2016 at 6:37 PM, Michael Paquier
wrote:
>
> On Sat, Jan 16, 2016 at 9:07 PM, Amit Kapila
wrote:
> > On Sat, Jan 16, 2016 at 5:08 PM, Michael Paquier <
michael.paqu...@gmail.com>
> > wrote:
> >>
> >> On Sat, Jan 16, 2016 at 7:10 PM, Amit
x27;s handler and validate into
am specific new file can make this arrangement closer to what
we have for PL's (ex. we have plpgsql_validator and plpgsql_call_
handler in pl_handler.c and similar handler and validator functions
for other languages in their corresponding modules).
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
On Mon, Jan 18, 2016 at 10:54 AM, Michael Paquier
wrote:
>
> On Sun, Jan 17, 2016 at 1:37 PM, Amit Kapila
wrote:
> > On Sat, Jan 16, 2016 at 6:37 PM, Michael Paquier <
michael.paqu...@gmail.com>
> > wrote:
> >>
> >>
> >>
> >
> >
On Mon, Jan 18, 2016 at 11:06 PM, Robert Haas wrote:
>
> On Mon, Jan 18, 2016 at 11:09 AM, Alvaro Herrera
> wrote:
> > Amit Kapila wrote:
> >
> >> The reason for not updating the patch related to this thread is that
it is
> >> dependent on the work for
nything relevant happened in any backend since the
> last time we performed a checkpoint/logged a running xacts snapshot.
>
Sounds to be a more robust way of dealing with this problem. Michael,
if you don't disagree with above proposal, then we can mark this patch
as Waiting on Author?
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
On Tue, Jan 19, 2016 at 12:41 PM, Michael Paquier
wrote:
>
> On Mon, Jan 18, 2016 at 10:19 PM, Amit Kapila
wrote:
> > On Mon, Jan 18, 2016 at 10:54 AM, Michael Paquier
> > wrote:
> >> Yes, the thing is that, as mentioned at the beginning of the thread, a
> >>
dvocating any such mechanism, but rather sharing an
information, I came across.
[1] - http://dsl.serc.iisc.ernet.in/publications/conference/bouquet.pdf
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
On Wed, Jan 20, 2016 at 12:41 AM, Robert Haas wrote:
>
> On Mon, Jan 18, 2016 at 10:41 PM, Amit Kapila
wrote:
> > Second important and somewhat related point is whether we should save
> > this information in PGPROC as 4 bytes or keep it in pgBackendStatus.
> > I think
7;t been reviewed at all, but the other two have seen a
> bit of discussion and evolution.
>
The last one "PGProc" has been reviewed and all the reported problems
have been fixed. It is "Ready For Committer".
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
minute that part of the cause for the
> > slowdown is slowness in pg_clog, maybe we should reconsider the initial
> > decision to flush as quickly as possible (i.e. adopt a strategy where
> > walwriter sleeps a bit between two flushes) in light of the group-update
> > feature
ess.
>
> Another idea popping to my mind in order to fix the CF manager burnout
> and actually motivate people into becoming CF managers: say that one a
> CF manager is done with a commit fest, folks on -hackers discuss if
> the CFM has done a good job or not. If yes, he gets a travel paid to
> the conference of his choice.
>
I also think there should be some way to give credit to CFM, if it is
difficult to do anything related to money, then we can enforce that if
CFM submits any patches for next CF, then those should be prioritised.
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
On Tue, Jan 12, 2016 at 2:41 PM, Dilip Kumar wrote:
> On Thu, Jan 7, 2016 at 4:53 PM, Andres Freund wrote:
>
>> On 2016-01-07 16:48:53 +0530, Amit Kapila wrote:
>>
>> I think it's a worthwhile approach to pursue. But until it actually
>> fixes the problem of l
isible to
any active transaction), it won't reduce the size which is already extended,
unless the empty space is at end of relation.
Are you updating any index column?
I think if you should once try with higher fill factor as you have lot
of updates.
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
L, wherever that occurs.
>
>
Thats a valid point and I think one way to retain such an evidence
without adding name of author/reviewer/committer is to add a link to
commit/'s after each feature, something like whats done in some of
the other release notes [1].
[1] - http://kerne
erstand why the recycle ratio is not 50%.
>
At the moment, I am also not able to see why it is so. You might
want to first try with a simple test (Can you extract insert/update
statements from application and run it manually for couple of times
and then run Vacuum to see the result).
By anychanc
On Sat, Jan 23, 2016 at 12:19 PM, Amit Kapila
wrote:
> On Tue, Jan 12, 2016 at 2:41 PM, Dilip Kumar
> wrote:
>
>> On Thu, Jan 7, 2016 at 4:53 PM, Andres Freund wrote:
>>
>>> On 2016-01-07 16:48:53 +0530, Amit Kapila wrote:
>>>
>>> I think it
On Wed, Jan 20, 2016 at 6:39 PM, Robert Haas wrote:
>
> On Tue, Jan 19, 2016 at 11:49 PM, Amit Kapila
wrote:
> >> On the topic of the UI, I understand that redefining
> >> pg_stat_activity.waiting might cause some short-term annoyance. But I
> >> think in th
e original Author even though the effort put by reviewer is
comparable to Author especially for somewhat complex patches.
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
On Tue, Jan 26, 2016 at 1:40 PM, and...@anarazel.de
wrote:
>
> On 2016-01-26 13:22:09 +0530, Amit Kapila wrote:
> > @@ -633,9 +633,11 @@ postgres 27093 0.0 0.0 30096 2752 ?
Ss 11:34 0:00 postgres: ser
> > Time when the state was last
changed
> >
>
formation
with other set of workers, then it will require such an enhancement.
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
ncept stage, rather than real
implementation, but I think it won't need too much effort to
improve it, if we find any particular approach as an acceptable
approach.
[1] -
http://www.postgresql.org/message-id/CA+TgmobWdBcbuipWPsbHSbf+-KDmatnYQYZ=akaju6alb5m...@mail.gmail.com
With Regard
here if existing users of pg_stat_activity.waiting are still
relying on it, they can get wrong information or if they are aware of the
recent change, then they need to add additional check like
(waiting = true && wait_event_type = 'HWLock')). So I think it is better to
go wi
On Thu, Jan 28, 2016 at 1:34 AM, james wrote:
> On 27/01/2016 13:30, Amit Kapila wrote:
>
>>
>> Thoughts?
>>
>> Are the decreases observed with SSD as well as spinning rust?
>
>
The test is done with WAL on SSD and data on spinning rust, but I
think the result
On Thu, Jan 28, 2016 at 8:07 PM, Magnus Hagander
wrote:
> Hello!
>
> The PostgreSQL core team would like to welcome Dean Rasheed as a new
> committer for the PostgreSQL project.
>
>
Many Congratulation Dean!
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
On Fri, Jan 29, 2016 at 6:21 PM, Alexander Korotkov <
a.korot...@postgrespro.ru> wrote:
> On Mon, Jan 4, 2016 at 5:58 PM, Robert Haas wrote:
>
>> On Mon, Jan 4, 2016 at 1:20 AM, Amit Kapila
>> wrote:
>> > If we do that way, then user of API needs to maintain
in the sentence below.
>
> + from _PG_init. Tranche repersents an array of LWLocks
>> and
>> + can be accessed by it's name. First parameter
>> tranche_name
>
>
>
Right, will change.
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
on:
Is there a reason why we should not assign ReplicationOrigins a
fixed tranche id and then we might want to even get away with
LWLockRegisterTranche()?
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
On Thu, Jan 28, 2016 at 9:16 AM, Amit Kapila
wrote:
>
> On Thu, Jan 28, 2016 at 2:12 AM, Robert Haas
wrote:
> >
> > On Tue, Jan 26, 2016 at 3:10 AM, and...@anarazel.de
wrote:
> > > I do think there's a considerable benefit in improving the
> > > in
On Sat, Jan 30, 2016 at 12:23 PM, Amit Kapila
wrote:
> On Fri, Jan 29, 2016 at 6:55 PM, Alexander Korotkov <
> a.korot...@postgrespro.ru> wrote:
>>
>> Also couple of minor comments from me.
>>
>> I think this
>>
>> + StrNCpy(LWLockTrancheReque
On Sat, Jan 30, 2016 at 2:15 PM, Alexander Korotkov <
a.korot...@postgrespro.ru> wrote:
>
> On Sat, Jan 30, 2016 at 10:58 AM, Amit Kapila
wrote:
>>
>> On Fri, Jan 29, 2016 at 6:47 PM, Robert Haas
wrote:
>> >
>> > On Fri, Jan 29, 2016 at 6:59 AM, Alexande
will be updated on
each XLOGInsert apart from the XLOGInsert for standby snapshots
and use that in a patch somewhat close to what you have in
hs-checkpoints-v1.
One related point is don't we need to bother about activity on
unlogged relations for checkpoints to occur, considering the
case when the only activity happened after last checkpoint
is on unlogged relations?
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
the second only one line.
>
Thank you!
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
On Mon, Feb 1, 2016 at 7:10 PM, Alexander Korotkov
wrote:
> On Sun, Jan 31, 2016 at 6:55 AM, Amit Kapila
> wrote:
>
>> On Thu, Jan 28, 2016 at 9:16 AM, Amit Kapila
>> wrote:
>> >
>> > On Thu, Jan 28, 2016 at 2:12 AM, Robert Haas
>> wrote:
>&
d) into two-bytes which sounds
reasonable to me apart from the fact that it might add operation or two
extra
in this path. Do you or anyone else have any preference over this point?
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
On Mon, Feb 1, 2016 at 8:05 PM, Jim Nasby wrote:
> On 1/31/16 3:26 PM, Jan Wieck wrote:
>
>> On 01/27/2016 08:30 AM, Amit Kapila wrote:
>>
>>> operation. Now why OS couldn't find the corresponding block in
>>> memory is that, while closing the WA
On Wed, Feb 3, 2016 at 11:12 AM, Amit Kapila
wrote:
>
> On Mon, Feb 1, 2016 at 8:05 PM, Jim Nasby
wrote:
>>
>> On 1/31/16 3:26 PM, Jan Wieck wrote:
>>>
>>> On 01/27/2016 08:30 AM, Amit Kapila wrote:
>>>>
>>>> operation. Now why OS
On Tue, Feb 2, 2016 at 7:19 PM, Robert Haas wrote:
>
> On Mon, Feb 1, 2016 at 12:27 AM, Amit Kapila
wrote:
> > Fixed.
>
> This patch doesn't build:
>
> ./xfunc.sgml:int lwlock_count = 0;
> Tabs appear in SGML/XML files
>
Changed
On Wed, Feb 3, 2016 at 3:08 PM, Andres Freund wrote:
>
> On 2016-02-02 10:12:25 +0530, Amit Kapila wrote:
> > @@ -8239,7 +8262,7 @@ CreateCheckPoint(int flags)
> > if ((flags & (CHECKPOINT_IS_SHUTDOWN |
> > CHECKPOINT_END_OF_RECOVERY |
> >CHECKPO
On Thu, Feb 4, 2016 at 6:40 PM, Andres Freund wrote:
>
> On 2016-02-04 18:21:41 +0530, Amit Kapila wrote:
> > I think generally it is good idea, but one thing worth a thought is that
> > by doing so, we need to acquire all WAL Insertion locks every
> > LOG_SNAPSHOT_
On Fri, Feb 5, 2016 at 3:17 AM, Robert Haas wrote:
>
> On Thu, Feb 4, 2016 at 7:00 AM, Amit Kapila
wrote:
> > [ new patch ]
>
> I've committed this after heavy rewriting. In particular, I merged
> two loops, used local variables to get rid of the very long and quite
&
e relation without checking if it can get a page with
space from FSM. It seems to me that we should re-check the
availability of page because while one backend is waiting on extension
lock, other backend might have added pages. To re-check the
availability we might want to use something similar to
LWLockAcquireOrWait() semantics as used during WAL writing.
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
ght not be full-proof, but OTOH we should
provide some way to allow user to start database and dump the
existing contents. Some of the options that comes to mind are
provide some way to get the last checkpoint record from WAL
or provide a way to compute max-lsn from data-pages and use
that with pg_resetxlog utility to allow user to start database.
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
On Sun, Feb 7, 2016 at 10:54 AM, Amit Kapila
wrote:
>
> On Tue, Feb 2, 2016 at 5:28 AM, Andres Freund wrote:
> >
> > Hi,
> >
> > currently if, when not in standby mode, we can't read a checkpoint
> > record, we automatically fall back to the previous ch
On Wed, Feb 3, 2016 at 7:12 PM, Robert Haas wrote:
>
> On Wed, Feb 3, 2016 at 7:28 AM, Amit Kapila
wrote:
> > On further testing, it has been observed that misaligned writes could
> > cause reads even when blocks related to file are not in-memory, so
> > I think what Ja
ten locked section
*/
>
> Check.
>
What is the need of holding locks one-by-one during checkpoint when
we anyway are going to take lock on all the insertion slots.
+ * to not rely on taking an exclusive lock an all the WAL insertion locks,
/an all/on all
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
On Mon, Feb 8, 2016 at 8:11 PM, Robert Haas wrote:
>
> On Mon, Feb 8, 2016 at 12:08 AM, Amit Kapila
wrote:
> > I think deciding it automatically without user require to configure it,
> > certainly has merits, but what about some cases where user can get
> > benefits b
On Mon, Feb 8, 2016 at 8:16 PM, Andres Freund wrote:
>
> On 2016-02-08 10:38:55 +0530, Amit Kapila wrote:
> > I think deciding it automatically without user require to configure it,
> > certainly has merits, but what about some cases where user can get
> > benefits by con
On Tue, Feb 9, 2016 at 5:37 AM, Michael Paquier
wrote:
>
> On Mon, Feb 8, 2016 at 11:24 PM, Amit Kapila
wrote:
> > On Mon, Feb 8, 2016 at 12:28 PM, Michael Paquier <
michael.paqu...@gmail.com>
> > wrote:
> >>
> >>
> >> >> /*
>
to me after new implementation
of named tranches, so I have removed that as well.
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
fixed_locks_tranche_v1.patch
Description: Binary data
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make chang
On Tue, Feb 9, 2016 at 6:08 PM, Michael Paquier
wrote:
>
> On Tue, Feb 9, 2016 at 2:24 PM, Amit Kapila
wrote:
> > Do you see any benefit in allowing checkpoints for such cases
considering
> > bgwriter will anyway take care of logging standby snapshot for such
> > cases?
On Wed, Feb 10, 2016 at 7:17 AM, Michael Paquier
wrote:
>
> On Tue, Feb 9, 2016 at 10:42 PM, Amit Kapila wrote:
> > On Tue, Feb 9, 2016 at 6:08 PM, Michael Paquier wrote:
> >> Well, the idea is to improve the system responsiveness. Imagine that
> >> the call to GetP
On Tue, Feb 9, 2016 at 7:26 PM, Thom Brown wrote:
>
> On 7 January 2016 at 05:24, Amit Kapila wrote:
> > On Fri, Dec 25, 2015 at 6:36 AM, Robert Haas
wrote:
> >>
> >> On Wed, Dec 23, 2015 at 1:16 AM, Amit Kapila
> >> wrote:
> >> >> >
On Tue, Feb 9, 2016 at 11:05 PM, Robert Haas wrote:
>
> On Tue, Feb 9, 2016 at 7:53 AM, Amit Kapila
wrote:
> > On Fri, Feb 5, 2016 at 3:17 AM, Robert Haas
wrote:
> >> I think we ought to move the buffer mapping, lock manager, and
> >> predicate lock manager lock
On Wed, Feb 10, 2016 at 12:16 PM, Michael Paquier wrote:
>
>
> On Wed, Feb 10, 2016 at 12:41 PM, Amit Kapila
> wrote:
> > On Wed, Feb 10, 2016 at 7:17 AM, Michael Paquier <
> michael.paqu...@gmail.com>
> > wrote:
> >>
> >
> > Consider below
On Wed, Feb 10, 2016 at 1:01 PM, Michael Paquier
wrote:
>
>
> On Wed, Feb 10, 2016 at 4:11 PM, Amit Kapila
> wrote:
>
>>
>>
>>
>>>
>>> >> > - last_snapshot_lsn != GetXLogInsertRecPtr())
>>> >> > +
>>> >>
On Wed, Feb 10, 2016 at 1:42 PM, Michael Paquier
wrote:
>
> On Wed, Feb 10, 2016 at 5:00 PM, Amit Kapila
wrote:
> > Okay, but isn't it better that we remove the snapshot taken
> > at checkpoint time in the main branch or till where this code is
> > getting back-patc
On Wed, Feb 10, 2016 at 8:51 PM, Robert Haas wrote:
>
> On Wed, Feb 10, 2016 at 1:32 AM, Amit Kapila
wrote:
> > The reason for using centralized way is that we need to request
> > named tranches before initialization of shared memory and as far as
> > I can see, currentl
On Thu, Feb 11, 2016 at 6:45 PM, Robert Haas wrote:
>
> On Thu, Feb 11, 2016 at 3:15 AM, Amit Kapila
wrote:
> > Sounds sensible, however after changes, CreateLWLocks() started
> > looking unreadable, so moved initialization and registration of tranches
> > to separate f
On Fri, Feb 12, 2016 at 12:39 AM, Robert Haas wrote:
>
> On Thu, Feb 11, 2016 at 12:10 PM, Amit Kapila
wrote:
> > Okay, changed as per suggestion.
>
> Looks good to me; committed.
>
Thanks!
Attached patch changes the name of some of the existing tranches
as suggested by
of multiple Clog-page updates in same group on the basis
that such cases will be less and even if it happens, it won't effect the
transaction status update.
Do you have anything else in mind?
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
se
+
RequestAddinLWLocks(1);
+#endif
+
+ RequestAddinShmemSpace(SHMEMMSGSZ);
+
It seems you have moved request for shared memory
(RequestAddinShmemSpace()) after requesting locks, any reason
for same?
You don't need to change the request for shared memory.
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
am expects some
data to be generated before-hand and the information of same is
added in file-header.
Thoughts?
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
prepared_stmt_parallel_query_v1.patch
Description: Binary data
/*
* Test case for PreparedStmt
* Need first to r
s.objid
AND blocking_locks.objsubid IS NOT DISTINCT FROM
blocked_locks.objsubid
AND blocking_locks.pid != blocked_locks.pid
JOIN pg_catalog.pg_stat_activity blocking_activity ON
blocking_activity.pid = blocking_locks.pid
WHERE NOT blocked_locks.GRANTED;
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
HASH_ELEM | HASH_BLOBS);
> }
>
>
Your proposed change seems right to me.
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
itional locking
has any impact on performance. I think taking locks at intervals
of 15s or 30s should not matter much, but it is better to be safe.
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
;t have any other outstanding timeouts than what
+ * this function
+ * uses. If
that stops being true, we could cancel the timeouts
+ * individually, but that'd be slower.
+ */
Comment seems to be misaligned.
3.
+ void
+ StandbyLockTimeoutHandler(void)
+ {
+ /* forget any pending STANDBY_DEADLOCK_TIMEOUT
On Sat, Feb 13, 2016 at 10:10 AM, Robert Haas wrote:
>
> On Fri, Feb 12, 2016 at 12:55 AM, Amit Kapila
wrote:
> > Very Good Catch. I think if we want to address this we can detect
> > the non-group leader transactions that tries to update the different
> > CLOG page (dif
On Sun, Feb 21, 2016 at 12:02 PM, Robert Haas wrote:
> On Sun, Feb 21, 2016 at 10:27 AM, Amit Kapila
> wrote:
>
>>
>> Client_Count/Patch_Ver 1 64 128 256
>> HEAD(481725c0) 963 28145 28593 26447
>> Patch-1 938 28152 31703 29402
>>
>>
>> We
On Sun, Feb 21, 2016 at 2:28 PM, Robert Haas wrote:
> On Sun, Feb 21, 2016 at 12:24 PM, Amit Kapila
> wrote:
>
>> On Sun, Feb 21, 2016 at 12:02 PM, Robert Haas
>> wrote:
>>
>>> On Sun, Feb 21, 2016 at 10:27 AM, Amit Kapila
>>> wrote:
>>>
&
On Wed, Feb 3, 2016 at 8:59 AM, Robert Haas wrote:
>
> On Tue, Feb 2, 2016 at 10:27 PM, Amit Kapila
wrote:
> > So, let's leave adding any additional column, but Alexander has brought
up
> > a good point about storing the wait_type and actual wait_event
> > informatio
On Wed, Feb 24, 2016 at 7:27 PM, Robert Haas wrote:
> On Wed, Feb 17, 2016 at 6:41 PM, Amit Kapila
> wrote:
> > Commit d1b7c1ffe72e86932b5395f29e006c3f503bc53d has added
> > the support for passing bind parameters to parallel workers, however
> > prepared statement wh
On Wed, Feb 24, 2016 at 7:14 PM, Robert Haas wrote:
> On Mon, Feb 22, 2016 at 10:05 AM, Amit Kapila
> wrote:
> >> I wouldn't bother tinkering with it at this point. The value isn't
> >> going to be recorded on disk anywhere, so it will be easy to change
> &
is as to why
the commit 6150a1b08a9fe7ead2b25240be46dddeae9d98e1 lead to degradation,
but any ideas are welcome.
[1] -
http://www.postgresql.org/message-id/CAB-SwXZh44_2ybvS5Z67p_CDz=XFn4hNAD=cnmef+qqkxwf...@mail.gmail.com
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
On Thu, Feb 25, 2016 at 11:38 PM, Simon Riggs wrote:
> On 24 February 2016 at 23:26, Amit Kapila wrote:
>
>> From past few weeks, we were facing some performance degradation in the
>> read-only performance bench marks in high-end machines. My colleague
>> Mithun, has t
On Fri, Feb 26, 2016 at 4:37 PM, Robert Haas wrote:
>
> On Thu, Feb 25, 2016 at 1:09 PM, Robert Haas
wrote:
> > On Thu, Feb 25, 2016 at 8:53 AM, Amit Kapila
wrote:
> >>> But if the user says
> >>> they want to PREPARE the query, they are probab
le file suggesting that adding more scripts to that group
>> would be bad. But yes, perhaps putting this test into its own standalone
>> group would be enough of a fix.
>
> Maybe now would be a good time to address this by applying the attached
> patch to master and seeing wh
xact_commit_timestamp
>
I think there is a value in exposing such a variant which takes bigint
and internally converts it to xid. I am not sure the semantics for
the other proposal txid_recent() is equivalent to what we have for
txid_current(). One thing which is different is that txid_curren
alog
information, they need to be aware of HeapTuple and other required
stuff like syscache? Again, if they need to update some stats or
something like that, they need to be aware of heap tuple format.
--
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
--
Sent via pgsql-ha
h difference in performance,
> small variance in perf can be attributed to variance in probability of
> drawing the particular built-in script.
>
>
Can you specify the m/c details as Andres wants tests to be conducted on
some high socket m/c?
--
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
ws huge benefit in this case.
> For sure, we should validate that it doesn't cause performance regression in
> other cases. At least we should test read-write and smaller machines.
> Any other ideas?
>
may be test on Power m/c as well.
--
With Regards,
Amit Kapila.
Enterp
On Fri, Aug 19, 2016 at 8:24 PM, Alexander Korotkov
wrote:
> On Fri, Aug 19, 2016 at 3:19 PM, Amit Kapila
> wrote:
>>
>> On Fri, Aug 19, 2016 at 11:42 AM, Alexander Korotkov
>> wrote:
>> > Hackers,
>> >
>> > originally this idea was prop
to use it.
[1] -
https://www.postgresql.org/message-id/caa4ek1lkq_udism-z2dq6cuvjh3db5fnfnnezzbpsrjw0ha...@mail.gmail.com
--
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
CPU intensive is probably a good bet to
push down to workers (provided it is safe).
--
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
On Sat, Aug 20, 2016 at 10:23 AM, Claudio Freire wrote:
> On Sat, Aug 20, 2016 at 1:05 AM, Amit Kapila wrote:
>> On Thu, Aug 18, 2016 at 8:24 AM, Claudio Freire
>> wrote:
>>>
>>> A couple of points make me uneasy about this patch, yet I can think of
>
should be set to false after dsm_impl_unpin_segment().
Do you need a check if (control_slot >= 0)? In the code just above
you have as Assert to ensure that it is >=0.
2.
+ if (dsm_control->item[seg->control_slot].pinned)
+ elog(ERROR, "cannot pin a segment that is already pinned&q
On Sat, Aug 20, 2016 at 9:58 PM, Claudio Freire wrote:
> On Sat, Aug 20, 2016 at 4:27 AM, Amit Kapila wrote:
>>
>> That makes sense, but this means there is a chance that the searches
>> could lead to different buffers in case of uniqueness checks (the
>> search wi
On Sat, Aug 20, 2016 at 6:13 PM, Robert Haas wrote:
> On Sat, Aug 20, 2016 at 7:37 AM, Amit Kapila wrote:
>> 2.
>> + if (dsm_control->item[seg->control_slot].pinned)
>> + elog(ERROR, "cannot pin a segment that is already pinned");
>>
>> Shou
On Mon, Aug 22, 2016 at 5:24 AM, Thomas Munro
wrote:
> On Sat, Aug 20, 2016 at 11:37 PM, Amit Kapila wrote:
>> On Tue, Aug 9, 2016 at 10:07 AM, Thomas Munro
>> wrote:
>>> On Tue, Aug 9, 2016 at 12:53 PM, Tom Lane wrote:
>>>> The larger picture here is that
sically
in similar situation, the count for nonExclusiveBackups will be
decremented and if it hits pg_start_backup_callback(), the following
Assertion can fail.
pg_start_backup_callback()
{
..
Assert(XLogCtl->Insert.nonExclusiveBackups > 0);
..
}
--
With Regards,
Amit Kapila.
EnterpriseDB: http://www.e
On Mon, Aug 22, 2016 at 6:06 PM, Thomas Munro
wrote:
> On Tue, Aug 23, 2016 at 12:07 AM, Amit Kapila wrote:
>> + int control_slot = -1;
>> ...
>> + if (control_slot == -1)
>> + elog(ERROR, "cannot unpin unknown segment handle");
>>
>> Isn't
On Tue, Aug 23, 2016 at 8:54 AM, Amit Kapila wrote:
> $SUBJECT will make hash indexes reliable and usable on standby.
> AFAIU, currently hash indexes are not recommended to be used in
> production mainly because they are not crash-safe and with this patch,
> I hope we can address tha
nager list as defined in
rmgrlist.h?
Another thing that needs some thoughts is the UI of this patch,
currently it is using integer mask which might not be best way, but
again as it is intended to be mainly used for tests, it might be okay.
Do we want to enable some tests in the regression
101 - 200 of 3368 matches
Mail list logo