Обсуждение: killing idle instances

Running PostgreSQL 7.2.1 on RedHat Linux 7.2.

A few nights ago, I was in the middle of building a new
table when I noticed that there were a lot of idle instances
and my table was waiting to be built (I believe the message
statement was WAITING or the like).

After considering the urgency of the tables' need to exist
before morning, I killed the idle instances and my process
began to function normally.

I am looking at the log file for PG and see a lot of  'BackendStartup:
forked pid=### socket=XXX'  messages and I wonder if I
did the right thing.

My question:  Should I *not* kill idle processes even though they
my interfere with building / modifying tables?  What are the
ramifications for such an act?

Thanks!

-X

On Thu, Apr 24, 2003 at 11:09:41AM -0400, Johnson, Shaunn wrote:

> A few nights ago, I was in the middle of building a new
> table when I noticed that there were a lot of idle instances
> and my table was waiting to be built (I believe the message
> statement was WAITING or the like).

The simple <idle> transactions wouldn't cause anyboy to wait.
Something must've been inside a transaction.

> I am looking at the log file for PG and see a lot of  'BackendStartup:
> forked pid=### socket=XXX'  messages and I wonder if I
> did the right thing.

That's the connections coming in.  Totally normal if you have the
right sort of logging options turned on.

> My question:  Should I *not* kill idle processes even though they
> my interfere with building / modifying tables?  What are the
> ramifications for such an act?

Truly idle (i.e. not in a transaction) processes shouldn't cause a
problem such as you describe.

A

--
----
Andrew Sullivan                         204-4141 Yonge Street
Liberty RMS                           Toronto, Ontario Canada
<andrew@libertyrms.info>                              M2P 2A8
                                         +1 416 646 3304 x110