Thread: LOG: could not receive data from client: Connection reset by peer

I'm getting the following errors on one of my pg server running postgres
7.4.5 on RedHat AS R4.

LOG:  could not receive data from client: Connection reset by peer
LOG:  incomplete startup packet

What I found out is that there's a networking VIP monitoring port 5432
every 15 sec and that's basically used for failover so that the
application will re-direct user requests to another server with
identical pg database in case of database or server failure.

I've checked that database and no connections are being created by the
VIP. Network admin says it operates on network layer 4 and just check if
the port is open. My question is, knowing all that, is there any caveat
or harm to Postgres implementing the VIP failover feature. Is there
anything I should be concerned of from a postgres perspective..

Thank you.


--
  Husam
**********************************************************************
This message contains confidential information intended only for the use of the addressee(s) named above and may
containinformation that is legally privileged.  If you are not the addressee, or the person responsible for delivering
itto the addressee, you are hereby notified that reading, disseminating, distributing or copying this message is
strictlyprohibited.  If you have received this message by mistake, please immediately notify us by replying to the
messageand delete the original message immediately thereafter. 

Thank you.

                                   FADLD Tag
**********************************************************************

"Tomeh, Husam" <htomeh@firstam.com> writes:
> I'm getting the following errors on one of my pg server running postgres
> 7.4.5 on RedHat AS R4.
> LOG:  could not receive data from client: Connection reset by peer
> LOG:  incomplete startup packet

> What I found out is that there's a networking VIP monitoring port 5432
> every 15 sec and that's basically used for failover so that the
> application will re-direct user requests to another server with
> identical pg database in case of database or server failure.

As long as you know what's causing it, it's safe to ignore --- it's not
harming anything except wasting a process launch each time.  If you
hadn't known the cause then I'd be worried about whether some hostile
is portscanning your network ...

What you *should* be worrying about is that you are still running 7.4.5.
That's two years old and there are a big pile of known bugs in it,
some of which *will* eat your data.

            regards, tom lane