Wed Mar 16 12:21:16 PST 2005
- Previous message: [Slony1-general] slony 1.1 release date
- Next message: [Slony1-general] slony 1.1 release date
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
?hel kenal p?eval (teisip?ev, 15. m?rts 2005, 12:34-0500), kirjutas Christopher Browne: > Indeed. > > There is a well-defined place to put tests; the "src/ducttape" directory > contains a set of shell scripts that test out a number of conditions. > > In the recent past, I have added in test_7_defines, which tests out the > new "include" and "define" slonik features, and test_8_logship, which > does some tests on the new log shipping functionality. > > Unfortunately, there are conditions which are difficult to set up tests > for. Any assistance in adding scripts to test particular failure cases > would be very much appreciated. Has anything been done to make slony *setup* less brittle ? I mean avoiding things like running a COPY for 48 hours, then failing to setup replica for a sequence, and merrily reporting 'sorry it failed, trying again in 15 sec'. This could be avoided if some sanity checks would be run before the real time-consuming actions when subscribing a set. These checks could be either separate functions, or perhaps a new no-op flag fro existing ones, so that two rounds of subscribe could be run, first with no_op=true and if it runs ok, only then a second with no_op=false to do the actual work. -- Hannu Krosing <hannu at skype.net>
- Previous message: [Slony1-general] slony 1.1 release date
- Next message: [Slony1-general] slony 1.1 release date
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Slony1-general mailing list