pinging reply from Equlizer delay

[kt]

:cry:
LB: e250si ver7.2.4g
Cluster config: single network / internal


Two e250si for failover configuration.

A. e250si(primary) Internal IP Address: 10.12.13.62
B. e250si(backup) Internal IP Address: 10.12.13.63
C. cluster IP Address : 10.12.13.64

From Windows Xp Client dos command prompt, using "ping" command,
Result is the below:
------------------------------------------------------------------------
> ping 10.12.13.64 -n 100

Pinging 10.12.13.64 with 32 bytes of data:

Reply from 10.12.13.64: bytes=32 time<1ms TTL=128
Reply from 10.12.13.64: bytes=32 time<1ms TTL=128
:
Reply from 10.12.13.64: bytes=32 time=530ms TTL=128 <== watch!!
:
Reply from 10.12.13.64: bytes=32 time<1ms TTL=128
------------------------------------------------------------------------
Replying from Equlizer always delay once every tens of times.
A, B also same result.

Thanks for your time and advise!

[gkalkat]

KT,

Thank you for contacting Coyote Point Support regarding your Equalizer product.

I'm not sure why there is random time delay values. I can only assume that the cluster is busy/under load processing requests.

Upon reviewing the information provided, we strongly recommend that you upgrade from version 7.2.4g to the latest version of software, which is 8.5.1d. There have been numerous bug fixes and new features added. In addition, you will notice better/faster performance along with a polished GUI.

Please refer to the attached document titled Upgrading Equalizer to Version 8.

Since there is a limit of one attachment per post, please retrieve the following documents from our Support Portal. Version 7 to Version 8 Transition Guide and the Release Notes for 8.5.1d.

- Gary Kalkat

[mark.hoffmann]

Just to be clear: I am assuming that this is an informational question and that you have not observed any processing issues on Equalizer connected with this behavior. Is that correct?

I also wanted to add a bit to what Gary said: running ping continually on a cluster IP is really not very meaningful in terms of response times, and probably bears absolutely no relationship to response times from servers behind the cluster. Pings are very low priority requests, so as Gary points out, if there is a lot of network traffic, then they may get delayed for higher priority traffic (like actual traffic to servers).

There is also the possibility that there is some daemon running on the 250si that will occasionally delay a ping response with no ill effects to load balancing processes. Since we have substantially updated the system (including probing and various daemons) since the release you are running, it may very well be that upgrading will just make this observed delay go away (assuming of course it is not due to the network configuration).

Hope this helps,

Mark