TCL Script to take rserver out of service

[shday]

I'm looking for some help with the LDAP_PROBE script that is included in the ACE30.  I want the script to monitor the LDAP server with the normal bind request thats included in the LDAP_PROBE, but when it fails I want the script to take it out of service so that it does not come backup unless I bring it back up.  In other words if the LDAP server doesn't respond to the bind request I want the configuration changed to take the rserver out of service.

Also, in need to make sure the rserver I take out of service is only the rserver that is supporting the 389 port for the bind request.  This rserver is in multiple farms, but only one farm using the 389 port.

[Kanwaljeet Singh]

Hi,

I am not sure how to modify the script but if you are using an existing script to monitor the real server and if the PROBE fails, ACE will automatically take that server out of "service" and you don't need anything to do. Ensure that PROBE script is associated with real server at the FARM level and not at the server level and that should also ensure that when probe fails, it only affects that server in that particular serverfarm only.

Let me know if you have any questions.

Regards,

Kanwal

Note: Please mark answers if they are helpful.

[shday]

Yes, I understand that the probe will take it out of service if the rserver fails to respond, but my issue is I don't want the server to be brought back up in service until I have time to run my diagnostic and i will manually bring the server back into service.  So what i need is the probe to take it out of service if it fails to respond to my probe then i need the rserver to be taken out of service in the configuration so it can't come back up without manual intervention.

[Kanwaljeet Singh]

Hi,

That is not possible with current capabilities of ACE. ACE will keep firing probes as per the interval configured and if server replies then according to pass count configured, it will make the server "operational". There is no way for ACE to stop firing probes once the server fails.

Regards,

Kanwal

Note: Please mark answers if they are helpful.

[shday]

So I can't modify the script to go into config mode and enter the no inservice command on the rserver if the probe doesn't get a succeful response?

[Kanwaljeet Singh]

Hi,

Again not sure about the script but on ACE (feature wise) that is not possible. If you are good with scripting you can try this but that's as far as i can help you with ACE:)

Regards,

Kanwal

Note: Please mark answers if they are helpful.

[Leon Breukelman]

Hello Shday, i have the same requirement and it seems that TCL script will work when you configure a scripted custom probe. On this URL:
http://www.cisco.com/c/en/us/td/docs/app_ntwk_services/data_center_app_services/ace_appliances/vA4_2_0/configuration/slb/guide/slbgd/script.html

there is a reference to an ACE spesific TCL command:

disable_real

and

enable_real

I am now trying to figure out how to put together such a TCL script, additionally, our dev team's requirement is that only 1 server in the SF may be active at any one moment, 

[Leon Breukelman]

Hello shday, have you made any progress on this?

I have found the following:

vsh_conf_cmd $cmd_string

Allows the execution of the command or set of commands specified in the preceding set command string (cmd_str) by invoking the Vegas shell (Vsh). If you specify more than one command in the command string, separate them by the \n characters. For example, enter: set cmd_str "rserver rs \n inservice" vsh_conf_cmd $cmd_str

Have you used this yet?

[shday]

I have used the TCL script and if I run the script as you have it in your reply it does take the rserver out of service, but I have not been able to get a script to monitor the status of the rserver and take it out of service if it fail and not automatically bring it back up.

I'm looking at using the inband probe instead as it will monitor the rserver and when the inband probe fails it take manually intervention to bring it back up.  I'm still testing this configuration.

serverfarm host name
  inband-health check remove 15 reset 300
  rserver name1 395
    backup-rserver name2 395
    inservice
  rserver name2 395
    inservice standby

The timers I have set here are just for testing.  We are still trying to determine the timers we would put in production if this configuration does what we want.