Thread: duplicated tables
Hi there, yesterday, testing GIN fast update with CVS HEAD I was able to crash backend (Teodor already fixed the problem in 0.25 version of the patch) and after restarting backend I found duplicated tables. How this can be possible and is this what somebody can see after crash ? List of relations Schema | Name | Type | Owner --------+-------+-------+---------- public | tt | table | postgres public | tt | table | postgres public | tt |table | postgres public | tt | table | postgres public | tt | table | postgres public | tt2 | table | postgres public| tt2 | table | postgres public | tt2 | table | postgres public | tt2 | table | postgres public | tt2 | table| postgres public | ttins | table | postgres public | ttins | table | postgres public | ttins | table | postgres public| ttins | table | postgres public | ttins | table | postgres (15 rows) Regards, Oleg _____________________________________________________________ Oleg Bartunov, Research Scientist, Head of AstroNet (www.astronet.ru), Sternberg Astronomical Institute, Moscow University, Russia Internet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/ phone: +007(495)939-16-83, +007(495)939-23-83
Oleg Bartunov <oleg@sai.msu.su> writes: > yesterday, testing GIN fast update with CVS HEAD I was able to crash backend > (Teodor already fixed the problem in 0.25 version of the patch) > and after restarting backend I found duplicated tables. > How this can be possible and is this what somebody can see after crash ? That doesn't look like duplicated tables so much as duplicated pg_namespace rows --- try doing \d's query by hand and adding display of pg_class.ctid and pg_namespace.ctid. However, if that theory is correct then the next question is what you were doing to pg_namespace... regards, tom lane
> That doesn't look like duplicated tables so much as duplicated > pg_namespace rows --- try doing \d's query by hand and adding display > of pg_class.ctid and pg_namespace.ctid. However, if that theory is > correct then the next question is what you were doing to pg_namespace... The bug was dereferencing uninitialized pointer, and postgres dumps core immediately. And patch does nothing with namespace. -- Teodor Sigaev E-mail: teodor@sigaev.ru WWW: http://www.sigaev.ru/
Teodor Sigaev escreveu: >> That doesn't look like duplicated tables so much as duplicated >> pg_namespace rows --- try doing \d's query by hand and adding display >> of pg_class.ctid and pg_namespace.ctid. However, if that theory is >> correct then the next question is what you were doing to pg_namespace... > > The bug was dereferencing uninitialized pointer, and postgres dumps core > immediately. And patch does nothing with namespace. > Could you post a backtrace or a test case? Maybe it is worth adding an assert or test at some point in the code. -- Euler Taveira de Oliveira http://www.timbira.com/