Re: Use incremental sort paths for window functions

2020-09-15 Thread David Rowley
On Tue, 15 Sep 2020 at 05:19, Tomas Vondra wrote: > > On Wed, Jul 08, 2020 at 04:57:21PM +1200, David Rowley wrote: > >Over on [1] someone was asking about chained window paths making use > >of already partially sorted input. (The thread is on -general, so I > >guessed they're not using PG13.) >

Re: Use incremental sort paths for window functions

2020-09-15 Thread Daniel Gustafsson
> On 15 Sep 2020, at 01:17, David Rowley wrote: > > On Tue, 15 Sep 2020 at 00:02, Daniel Gustafsson wrote: >> >>> On 8 Jul 2020, at 06:57, David Rowley wrote: >>> >>> Over on [1] someone was asking about chained window paths making use >>> of already partially sorted input. (The thread is

Re: New statistics for tuning WAL buffer size

2020-09-15 Thread Fujii Masao
On 2020/09/15 15:52, Masahiro Ikeda wrote: On 2020-09-11 01:40, Fujii Masao wrote: On 2020/09/09 13:57, Masahiro Ikeda wrote: On 2020-09-07 16:19, Fujii Masao wrote: On 2020/09/07 9:58, Masahiro Ikeda wrote: Thanks for the review and advice! On 2020-09-03 16:05, Fujii Masao wrote: On

Re: Parallelize stream replication process

2020-09-15 Thread Fujii Masao
On 2020/09/15 13:41, Bharath Rupireddy wrote: On Tue, Sep 15, 2020 at 9:27 AM Li Japin wrote: For now, postgres use single process to send, receive and replay the WAL when we use stream replication, is there any point to parallelize this process? If it does, how do we start? Any

Re: pg_restore causing deadlocks on partitioned tables

2020-09-15 Thread Amit Langote
On Tue, Sep 15, 2020 at 9:09 AM Tom Lane wrote: > I wrote: > >> (2) ALTER TABLE ONLY ... ADD CONSTRAINT on a partition root tries to get > >> AccessExclusiveLock on all child partitions, despite the ONLY. > > > The cause of this seems to be that ATPrepSetNotNull is too dumb to > > avoid recursing

Re: Allow ERROR from heap_prepare_freeze_tuple to be downgraded to WARNING

2020-09-15 Thread Dilip Kumar
On Tue, Sep 15, 2020 at 11:14 AM Andres Freund wrote: > > On 2020-09-15 10:54:29 +0530, Dilip Kumar wrote: > > What problem do you see if we set xmax to the InvalidTransactionId and > > HEAP_XMAX_INVALID flag in the infomask ? > > 1) It'll make a dead tuple appear live. You cannot do this for

Re: logtape.c stats don't account for unused "prefetched" block numbers

2020-09-15 Thread Peter Geoghegan
On Mon, Sep 14, 2020 at 11:52 PM Jeff Davis wrote: > On Thu, 2020-09-10 at 18:42 -0700, Peter Geoghegan wrote: > > We still need to put the reliance on ltsWriteBlock() allocating many > > blocks before they've been logically written on some kind of formal > > footing for Postgres 13 -- it is now

Re: New statistics for tuning WAL buffer size

2020-09-15 Thread Masahiro Ikeda
On 2020-09-11 01:40, Fujii Masao wrote: On 2020/09/09 13:57, Masahiro Ikeda wrote: On 2020-09-07 16:19, Fujii Masao wrote: On 2020/09/07 9:58, Masahiro Ikeda wrote: Thanks for the review and advice! On 2020-09-03 16:05, Fujii Masao wrote: On 2020/09/02 18:56, Masahiro Ikeda wrote: +/*

Re: logtape.c stats don't account for unused "prefetched" block numbers

2020-09-15 Thread Jeff Davis
On Thu, 2020-09-10 at 18:42 -0700, Peter Geoghegan wrote: > We still need to put the reliance on ltsWriteBlock() allocating many > blocks before they've been logically written on some kind of formal > footing for Postgres 13 -- it is now possible that an all-zero block > will be left behind even

Re: logtape.c stats don't account for unused "prefetched" block numbers

2020-09-15 Thread Jeff Davis
On Mon, 2020-09-14 at 20:48 -0700, Peter Geoghegan wrote: > On Mon, Sep 14, 2020 at 8:07 PM Jeff Davis wrote: > > Sure. Will backporting either patch into REL_13_STABLE now > > interfere > > with RC1 release in any way? > > The RMT will discuss this. > > It would help if there was a patch ready

<    1   2