Re: Is TNSPING = 1000ms OK? Oracle says it is

  • From: Mladen Gogala <gogala@xxxxxxxxxxxxx>
  • To: Tony Jambu <tjambu_freelists@xxxxxxxxxxxx>
  • Date: Wed, 08 Mar 2006 00:45:26 -0500

On 03/07/2006 04:44:21 PM, Tony Jambu wrote:
> Hi Mladen
> Will do some reading on it.  But in the meantime,
> why is it that both TNSPING to the same server at the same
> time (to different ports) return different times?  Who is the one
> that initial the 'type' ping.  The initiator or 'target'

Tony, 1 second for the packet to come back is far too much. My first 
advice would be to try normal unix ping. That sould give you
the average packet turnaround time between two nodes. Values
reported by tnsping should be somewhat larger then the values
reported by the normal "ping" utility. What TNSPING does is to
send packets to the listener and listener responds. Normal, unix
ping uses ICMP protocol to elicit response. It doesn't go through
TCP layer. Values up to twice as large as ping are expected. 
Values 10 times larger show you that the anomaly is in the
oracle layer. Next thing to do is to test with nmap utility.
This utility is not present at every system and you might have
to ask your admin to install it. It's one among the best hacker
tools, normally used for searching open ports. In contrast to
"ping", "nmap" knows how to use TCP and will ping the port
and time the ping. If both ports are reporting the same time
to nmap, you can do only one thing: trace the connection from
both sides and get Oracle involved. 

Mladen Gogala


Other related posts: