slow ping

[Eyal Teutsch]

Hi,

I saw a couple of issues dealing with this problem, but could not really figure out how to solve/diagnose my problem further.

We're managing our VM centre using VMware Lab Manager which runs on a VMware ESX Server.

In some cases, when we execute a STAF ping command from one of our desktop (linux, non VM) to a vm host in the lab (linux too), we get a slow response (~5 seconds, as opposed to a normal ~0.15 sec we normally get).

The problem is usually quite consistent, but I also saw it solved once by simply restarting STAF.

regular ping to these hosts work without any problem.

thx,

Eyal

[David Michael Bender]

What versions of STAF are you running on the systems (desktop and vm host)?  When the problem occurs, what is the output of the following (run on the desktop):

STAF local MISC LIST INTERFACES

STAF <vm-host> MISC LIST INTERFACES

[Eyal Teutsch]

Hi,

any thoughts/insights on the output given?

thx,

eyal

[Eyal Teutsch]

on our linux desktop machines we're running 3.3.3, on the vms we're running either 3.3.2 or 3.3.3 (we have the problem in both cases).

local interfaces output:
$ STAF local MISC LIST INTERFACES
Response
--------
[
  {
    Interface Name: local
    Library       : STAFLIPC
    Options       : {
      IPCName: STAF
    }
  }
  {
    Interface Name: ssl
    Library       : STAFTCP
    Options       : {
      ConnectTimeout       : 5000
      Port                 : 6550
      Protocol             : IPv4
      SSL/CACertificate    : /usr/local/staf/bin/CAList.crt
      SSL/ServerCertificate: /usr/local/staf/bin/STAFDefault.crt
      SSL/ServerKey        : /usr/local/staf/bin/STAFDefault.key
      Secure               : Yes
    }
  }
  {
    Interface Name: tcp
    Library       : STAFTCP
    Options       : {
      ConnectTimeout: 5000
      Port          : 6500
      Protocol      : IPv4
      Secure        : No
    }
  }
]

vm-host output:
$ STAF 10.9.7.245 MISC LIST INTERFACES
Response
--------
[
  {
    Interface Name: local
    Library       : STAFLIPC
    Options       : {
      IPCName: STAF
    }
  }
  {
    Interface Name: ssl
    Library       : STAFTCP
    Options       : {
      ConnectTimeout       : 5000
      Port                 : 6550
      Protocol             : IPv4
      SSL/CACertificate    : /usr/local/staf/bin/CAList.crt
      SSL/ServerCertificate: /usr/local/staf/bin/STAFDefault.crt
      SSL/ServerKey        : /usr/local/staf/bin/STAFDefault.key
      Secure               : Yes
    }
  }
  {
    Interface Name: tcp
    Library       : STAFTCP
    Options       : {
      ConnectTimeout: 5000
      Port          : 6500
      Protocol      : IPv4
      Secure        : No
    }
  }
]

[David Michael Bender]

Have you tried any of the suggestions in the following forum threads:

http://sourceforge.net/forum/message.php?msg_id=4174813
http://sourceforge.net/forum/message.php?msg_id=2763377

[Eyal Teutsch]

I had a look at those topics, and may have missed something, but it seemed to me that suggestions there relate more to a situation where you have a windows machine involved, while in our case it's linux on both ends (and btw - we're pinging the IP rather than the hostname).

[David Michael Bender]

If you do a non-ping remote connection between the systems, do you get the same 5-second response?  For example, from one Linux system start a ssh/telnet session to the other Linux system.  Does the ssh/telnet request respond immediately or take 5 seconds?

[Eyal Teutsch]

the normal ping is immediate - no delay occurs there.

[David Michael Bender]

Right, the normal ping is immediate.  What I want to know is if other methods of remote access to the system are immediate.

As an example, I have some Linux systems at a different site.  If I try to do a normal ping to these systems (from my local site), the response is immediate.  If I do a STAF PING to these systems, however, there is a 10-second delay for the response.  However, if I also run "telnet <system>" (from my local site), it also takes 10-seconds for the userid prompt to be displayed.  So, clearly, the delay has something to do with the network configuration, and is not an issue specific to STAF.  I am trying to understand if that is the same situation in your environment.

[shiva]

I had a similar issue, my ping request used to take more than 15 sec to respond. Turned out to be a DNS resolution issue. Either have a DNS server in your domain or edit the /etc/hosts file to resolve the host names in both client and server machines. Hope that helps!