Re: snapshot too old, configured by time
От | Thom Brown |
---|---|
Тема | Re: snapshot too old, configured by time |
Дата | |
Msg-id | CAA-aLv5r3QXCU44qLt88cTcnGOqGxPTsLBM7S-pndkQ3RmG=yA@mail.gmail.com обсуждение исходный текст |
Ответ на | Re: snapshot too old, configured by time (Kevin Grittner <kgrittn@gmail.com>) |
Ответы |
Re: snapshot too old, configured by time
|
Список | pgsql-hackers |
On 17 March 2016 at 21:15, Kevin Grittner <kgrittn@gmail.com> wrote: > New patch just to merge in recent commits -- it was starting to > show some bit-rot. Tests folded in with main patch. In session 1, I've run: # begin transaction isolation level repeatable read ; BEGIN *# declare stuff scroll cursor for select * from test where num between 5 and 9; DECLARE CURSOR *# fetch forward 5 from stuff;id | num | thing -----+-----+------------------------------------ 2 | 8 | hellofji djf odsjfiojdsif ojdsiof 3 | 7 | hellofji djf odsjfiojdsifojdsiof112 | 9 | hellofji djf odsjfiojdsif ojdsiof115 | 6 | hellofji djf odsjfiojdsif ojdsiof119 | 8 |hellofji djf odsjfiojdsif ojdsiof (5 rows) In session 2, over a min later: # update test set num = 12 where num between 5 and 9 and id between 120 and 250; ERROR: snapshot too old Then back to session 1: *# fetch forward 5 from stuff; ERROR: snapshot too old Should session 2 be getting that error? Thom
В списке pgsql-hackers по дате отправления: