Re: [HACKERS] Declarative partitioning optimization for large amount of partitions

2017-03-24 Thread Amit Langote
On Fri, Mar 24, 2017 at 11:06 PM, Simon Riggs wrote: > On 1 March 2017 at 01:36, Amit Langote wrote: > >> I don't know which way you're thinking of fixing this, but a planner patch >> to implement faster partition-pruning will have taken care

Re: [HACKERS] Declarative partitioning optimization for large amount of partitions

2017-03-24 Thread Aleksander Alekseev
Hi Simon, > > I don't know which way you're thinking of fixing this, but a planner patch > > to implement faster partition-pruning will have taken care of this, I > > think. As you may know, even declarative partitioned tables currently > > depend on constraint exclusion for partition-pruning

Re: [HACKERS] Declarative partitioning optimization for large amount of partitions

2017-03-24 Thread Simon Riggs
On 1 March 2017 at 01:36, Amit Langote wrote: > I don't know which way you're thinking of fixing this, but a planner patch > to implement faster partition-pruning will have taken care of this, I > think. As you may know, even declarative partitioned tables

Re: [HACKERS] Declarative partitioning optimization for large amount of partitions

2017-03-10 Thread Aleksander Alekseev
Hi Tels, Thanks a lot for the review! > "corresponding" Fixed. > Also a question: Some one-line comments are > > /* Comment. */ > > while others are > > /* > * Comment. > */ > > Why the difference? I'm trying to follow a code stile used in a code I'm modifying. In this case I got an

Re: [HACKERS] Declarative partitioning optimization for large amount of partitions

2017-03-09 Thread Tels
Hi Aleksander, noticed this in your patch: > * Add a corespinding entry to pgStatTabHash. "corresponding" Also a question: Some one-line comments are /* Comment. */ while others are /* * Comment. */ Why the difference? Hope this helps, Tels PS: Sorry if this appears twice, I

Re: [HACKERS] Declarative partitioning optimization for large amount of partitions

2017-03-09 Thread Tels
Hi Aleksander, noticed this in your patch: > * Add a corespinding entry to pgStatTabHash. "corresponding" Also a question: Some one-line comments are /* Comment. */ while others are /* * Comment. */ Why the difference? Hope this helps, Tels -- Sent via pgsql-hackers mailing list

Re: [HACKERS] Declarative partitioning optimization for large amount of partitions

2017-03-09 Thread Aleksander Alekseev
Hi Amit, > Sorry, I didn't mean to dissuade you from trying those > micro-optimizations. If general inheritance cases can benefit from it > (which, until we have a different method, will be used by partitioned > tables as well), I think we should try it. OK, I'll see what could be done here as

Re: [HACKERS] Declarative partitioning optimization for large amount of partitions

2017-03-09 Thread Aleksander Alekseev
Hi, Andres Thanks a lot for the review! > Why are we keeping that list / the "batch" allocation pattern? It > doesn't actually seem useful to me after these changes. Given that we > don't shrink hash-tables, we could just use the hash-table memory for > this, no? I don't think we can do that.

Re: [HACKERS] Declarative partitioning optimization for large amount of partitions

2017-03-06 Thread Amit Langote
Hi Aleksander, On 2017/03/07 0:22, Aleksander Alekseev wrote: > Hello. > > OK, here is a patch. > > Benchmark, before: > > ``` > number of transactions actually processed: 1823 > latency average = 1153.495 ms > latency stddev = 154.366 ms > tps = 6.061104 (including connections establishing) >

Re: [HACKERS] Declarative partitioning optimization for large amount of partitions

2017-03-06 Thread Andres Freund
Hi, This issue has bothered me in non-partitioned use-cases recently, so thanks for taking it up. On 2017-03-06 18:22:17 +0300, Aleksander Alekseev wrote: > diff --git a/src/backend/postmaster/pgstat.c b/src/backend/postmaster/pgstat.c > index 2fb9a8bf58..fa906e7930 100644 > ---

Re: [HACKERS] Declarative partitioning optimization for large amount of partitions

2017-03-06 Thread Aleksander Alekseev
Hello. OK, here is a patch. Benchmark, before: ``` number of transactions actually processed: 1823 latency average = 1153.495 ms latency stddev = 154.366 ms tps = 6.061104 (including connections establishing) tps = 6.061211 (excluding connections establishing) ``` Benchmark, after: ``` number

Re: [HACKERS] Declarative partitioning optimization for large amount of partitions

2017-02-28 Thread Amit Langote
Hi, On 2017/02/28 23:25, Aleksander Alekseev wrote: > Hello. > > I decided to figure out whether current implementation of declarative > partitioning has any bottlenecks when there is a lot of partitions. Here > is what I did [1]. Thanks for sharing. > Then: > > ``` > # 2580 is some pk that