> In http://slony.info/documentation/requirements.html#TIMES it says:
>
> Those timezones are sure to be supported on any platform, and have the
> merit over "local" timezones that times never wind up leaping around due
> to Daylight Savings Time.
>
> Just how bad is the daylight savings time switch for slony?  Any why is
> it a problem in the first place?  It should be easy for slony to get the
> current UTC time.

I wouldn't expect big problems to fall out of the timezone switch, even in
a context where you're doing DST.

Slony-I operates based on counting which events have been applied from
each source, and that isn't based on any comparison of time intervals.

That is, when a slon is figuring out what work to do, it'll essentially say:

"Hmm.  For node #2, I have processed up to event #4278.  I see node #2
publishing events #4279 thru #4285.  I guess I'd better process them."

And notice that that sentence never said anything about what timestamps or
time intervals were associated with anything.

I can think of one case where time intervals *are* considered, and that's
if you use the slon -l option which causes it to lag by some period of
time.  That lag time might either disappear or get extended by an hour
when a DST transition takes place.  I expect that -l is a very
infrequently used option, though...

_______________________________________________
Slony1-general mailing list
[email protected]
http://gborg.postgresql.org/mailman/listinfo/slony1-general

Reply via email to