"Robert A. Schmied" wrote yesterday before doing all the research: > > maurice et al > > find the attached diff -bitwuB for subject module test.pl.orig > and my hack of it in order to get all tests to pass on my > ancient ss20. > > the basic problem was the system refused to accept any -size value > less than 200 (or zero) in the new method. my hack forces it to > 200 with a factor to ensure tests 5, 6 and 7 all pass. > > aloha > > ras > << big snip >>
humm, well i did a bit more digging and found the source of the shared memory minimum . set via /etc/system to 200. done according to note in file per sol2.6 admin guide (vol 2 pg 1042 ) for postgresql. a later note from postgresql 8.0.3 doc/html/kernel-resources.html suggests using set shmsys:shminfo_shmmin=1 which has yet to be enabled. now if i can figure a way to access these system variables via perl ... so unless you think my additional failure message strings are significant feel free to put my patch list in (or is it on) /dev/null. aloha ras
