Martijn van Oosterhout <kleptog@svana.org> writes:
> On Mon, Mar 12, 2007 at 01:13:42PM -0600, Ed L. wrote:
>> Would I be correct in understanding that every pre-8.0 cluster
>> must be restarted in order for the OS changes to take affect?!?
> Possibly, I imagine many C libraries would cache the timezone data over
> a fork and might not motice the changes...
Postgres is hardly the only app that behaves like that. I saw a bunch
of traffic on the Red Hat lists yesterday/today about cron, syslogd,
etc needing restarts in various releases.
regards, tom lane