On 2012-11-18 13:18:42 -0500, Tom Lane wrote:
> Andres Freund <andres@2ndquadrant.com> writes:
> > On 2012-11-18 12:44:55 -0500, Tom Lane wrote:
> >> I'm inclined to propose limiting both of these to the equivalent of 15
> >> days. Alternatively we could try to rejigger things to prevent asking
> >> WaitLatch to wait for more than 2^31 msec, but it's not real clear to
> >> me that it's worth the trouble.
>
> > In general I have no problem imposing lower limits, but it seems to be a
> > ugly to get errors for an invalid configuration file after a minor
> > version upgrade. While the wal_sender_timeout isn't really likely to be
> > that high I don't think the log_rotation_age one is unlikely to be set
> > to something in the month range, thats not an unreasonable value.
>
> Well, we have two reports of people trying such values (assuming that
> #7545 actually is the same thing), and it didn't work for either of
> them. I don't think it's a problem to restrict the value to something
> that will work rather than fail.
Thats a good point. But #7545 was on OSX so its not really that much
evidence a larger value doesn't work on non-bsdish systems.
In fact, a setting of 32 days seems not to cause any immediate problems
here on linux, even when choosing the timeout in a way it results in a
negative timeout value for poll. Not sure what its waiting for, but it
doesn't crash.
Greetings,
Andres Freund