I've asked russian postgresql users to proceed my deadlock example. http://sql.ru/forum/actualthread.aspx?tid=418296
Here is the results. OS, version, build PG version, build Status (Person) SUSE Linux 10.1 8.1.4 from official SUSE distribution deadlock between /*1*/ and /*3*/ detected (Nalbat) SUSE Linux 10.1 8.2.3 self-made package /*3*/ waits for /*1*/, deadlock not detected (Nalbat) itanium2(Debian) 8.2.3 deadlock not detected (SergeyK) laptop(Slackware) 8.2.3 deadlock not detected (SergeyK) SUSE 10.0 8.1.4 self-made deadlock detected (SergeyK) Windows 2000 Pro Rus 8.0.4 /*2*/ waits, deadlock not detected (ZemA) Windows 2000 Pro Rus 8.1.3 deadlock detected (ZemA) win xp sp2 8.2.0, i686-pc-mingw32, compiled by GCC gcc.exe (GCC) 3.4.2 (mingw-special) /*3*/ waits for /*1*/, deadlock not detected (st_serg) It seems that deadlock rises for Postgresql 8.1 at any operating system. Tom Lane wrote: > Alexey Nalbat <[EMAIL PROTECTED]> writes: > > Tom Lane wrote: > >> Your example doesn't deadlock for me ... > > > With default value "deadlock_timeout=1000" error raises in first > > transaction: > > Then there's a typo in your example --- please recheck it. -- Alexey A. Nalbat Price Express http://www.price.ru/ http://www.tyndex.ru/ ---------------------------(end of broadcast)--------------------------- TIP 9: In versions below 8.0, the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
