Re: Insert into partition table hangs

"Nik" <XLPizza(at)gmail(dot)com> wrote
>
> I also noticed that I am having the similar issue with SELECT
> statements. Same SELECT statement will run fine 3-4 times in a row and
> then get stuck and run forever with no results. But if I reload the
> configuration like above, the query immediatelly returns with correct
> result.
>
> Why does this happen, and could this give any insight in my problem?
>

Interesting enough :-(

I can see that for some reason, the query may be blocked on a semaphore and
also listen to the signal. So when the SIGHUP comes, it will wake up and set
a variable "got_SIGHUP" to true (that's the only immediate thing do for
SIGHUP). After that, we either continue to sleep on the semaphore or exit
with a FATAL message -- both situations does not match your senario.

A wild guess is that when we continue to sleep on the semaphore, magically
we get it, but how come?

Regards,
Qingqing