Re: [HACKERS] Vacuuming big btree indexes without pages with deleted items

2015-04-01 Thread Vladimir Borodin
31 марта 2015 г., в 23:33, Kevin Grittner kgri...@ymail.com написал(а): Jim Nasby jim.na...@bluetreble.com wrote: On 3/27/15 5:15 AM, Vladimir Borodin wrote: Master writes this record to xlog in btvacuumscan function after vacuuming of all index pages. And in case of no pages with

Re: [HACKERS] Vacuuming big btree indexes without pages with deleted items

2015-03-31 Thread Kevin Grittner
Jim Nasby jim.na...@bluetreble.com wrote: On 3/27/15 5:15 AM, Vladimir Borodin wrote: Master writes this record to xlog in btvacuumscan function after vacuuming of all index pages. And in case of no pages with deleted items xlog record would contain lastBlockVacuumed 0. In btree_xlog_vacuum

Re: [HACKERS] Vacuuming big btree indexes without pages with deleted items

2015-03-30 Thread Jim Nasby
On 3/27/15 5:15 AM, Vladimir Borodin wrote: Hi all. I have described [0] a problem with delaying replicas after vacuuming a relation with big btree index. It stucks in replaying WAL record of type XLOG_BTREE_VACUUM like that (with lastBlockVacuumed 0): rmgr: Btree len (rec/tot): 20/

[HACKERS] Vacuuming big btree indexes without pages with deleted items

2015-03-27 Thread Vladimir Borodin
Hi all. I have described [0] a problem with delaying replicas after vacuuming a relation with big btree index. It stucks in replaying WAL record of type XLOG_BTREE_VACUUM like that (with lastBlockVacuumed 0): rmgr: Btree len (rec/tot): 20/52, tx: 0, lsn: 4115/56126DC0,