Re: [HACKERS] New gist vacuum.

2015-12-01 Thread Костя Кузнецов
Thank you, Jeff.I reworking patch now. All // warning will be deleted.About memory consumption new version will control size of stack and will operate with map of little size because i want delete old style vacuum(now if maintenance_work_mem less than needed to build info map we use old-style vacuu

[HACKERS] New gist vacuum.

2015-09-11 Thread Костя Кузнецов
Hello. I am student from gsoc programm.My project is sequantial access in vacuum of gist. New vacuum has 2 big step:physical order scan pages and cleaning after 1 step. 1 scan - scan all pages and create information map(hashmap) and add information to rescan stack( stack of pages that needed to re

Re: [HACKERS] gist vacuum seaq access

2014-09-20 Thread Костя Кузнецов
Heikki.I have idea. when i begining vacuum i must create structure like hash table. And the first action i fill this hash table.in code this look like:for( blkno = ..; all pages; blkno++) {  if(! gistPageIsLeaf(blkno)) { for( all tuples in blkno ) { hash[ blkno that is referenced by the

[HACKERS] gist vacuum seaq access

2014-09-11 Thread Костя Кузнецов
After discussion of gist seaq access in vaccum there are 2 issue: Heikki says :Vacuum needs to memorize the current NSN when it begins1) how i may getting corect NSN. Also i must setting F_DELETED flag on empty page and to clean parent from link on deleted_pages2) how i may getting parent of page f

[HACKERS] gist vacuum gist access

2014-09-07 Thread Костя Кузнецов
hello. i recode vacuum for gist index.all tests is ok.also i test vacuum on table size 2 million rows. all is ok.on my machine old vaccum work about 9 second. this version work about 6-7 sec .review please. thanks.*** a/src/backend/access/gist/gistvacuum.c --- b/src/backend/access/gist/gistvacuum.c

[HACKERS] Sequential disk access during VACUUM for GiST/GIN

2014-04-30 Thread Костя Кузнецов
Hello.There is a task "Sequential disk access during VACUUM for GiST/GIN " in list GSOC14.Nobody is working on this task?Do I understand this task correctly?I must recode gistbulkdelete.GistBDItem *stack is must have items with sequential blkno as possible.I have a question:where are access to disk

[HACKERS] gsoc knn spgist

2014-04-04 Thread Костя Кузнецов
Helllo. I want to implement knn for spgist. I dont have question with knn, but have questions with implementation of interface. i modify pg_am.h (set amcanorderbyop  to true in spgist index).Also i modify pg_amop.h(add  DATA(insert (    4015   600 600 15 o 517 4000 1970 )); ) explain SELECT * FROM

[HACKERS] gsoc knn spgist

2014-03-25 Thread Костя Кузнецов
Hello. I submit a proposal. But Heikki Linnakangas write comments that i dont have a plan of implementation. My project is knn for spgist. Can I ask you a few questions? 1. I research a commit gist knn implementation. in gist implementation in role of queue is ised rtree(with distance comparator) ,