Error creating virtual server

[andrew80k]

Creating a new virtual server I'm getting an odd error that I can't seem to track down.

This particular test playbook creates a set of servers, a service group, binds the servers to the service group, then creates the virtual server with the service group bound to it. This error only occurs on the initial creation run. Subsequent runs finish successfully. Second run has a single change in the vserver and any additional runs of the playbook remain unchanged.

fatal: [localhost -> localhost]: FAILED! => {
    "changed": true, 
    "diff": {
        "timeout": "difference. ours: (<type 'float'>) 10.0 other: (<type 'int'>) 2"
    }, 
    "invocation": {
        "module_args": {
            "appflowlog": null, 
            "appfw_policybindings": null, 
            "authentication": null, 
            "authenticationhost": null, 
            "authn401": null, 
            "authnprofile": null, 
            "authnvsname": null, 
            "backuplbmethod": null, 
            "backuppersistencetimeout": null, 
            "bypassaaaa": null, 
            "cacheable": null, 
            "clttimeout": null, 
            "comment": null, 
            "connfailover": null, 
            "cookiename": null, 
            "datalength": null, 
            "dataoffset": null, 
            "dbprofilename": null, 
            "dbslb": null, 
            "disabled": false, 
            "disableprimaryondown": null, 
            "dns64": null, 
            "dnsprofilename": null, 
            "downstateflush": null, 
            "hashlength": null, 
            "healththreshold": null, 
            "httpprofilename": null, 
            "icmpvsrresponse": null, 
            "insertvserveripport": null, 
            "instance_ip": null, 
            "ipmask": null, 
            "ippattern": null, 
            "ipv46": "10.81.244.79", 
            "l2conn": null, 
            "lbmethod": null, 
            "listenpolicy": null, 
            "listenpriority": null, 
            "m": null, 
            "macmoderetainvlan": null, 
            "mas_proxy_call": false, 
            "maxautoscalemembers": null, 
            "minautoscalemembers": null, 
            "mssqlserverversion": null, 
            "mysqlcharacterset": null, 
            "mysqlprotocolversion": null, 
            "mysqlservercapabilities": null, 
            "mysqlserverversion": null, 
            "name": "gen-anstest", 
            "netmask": null, 
            "netprofile": null, 
            "newservicerequest": null, 
            "newservicerequestincrementinterval": null, 
            "newservicerequestunit": null, 
            "nitro_auth_token": null, 
            "nitro_pass": "VALUE_SPECIFIED_IN_NO_LOG_PARAMETER", 
            "nitro_protocol": "https", 
            "nitro_timeout": 310.0, 
            "nitro_user": "VALUE_SPECIFIED_IN_NO_LOG_PARAMETER", 
            "nsip": "10.81.228.156", 
            "oracleserverversion": null, 
            "persistencebackup": null, 
            "persistencetype": null, 
            "persistmask": null, 
            "port": 80, 
            "processlocal": null, 
            "push": null, 
            "pushlabel": null, 
            "pushmulticlients": null, 
            "pushvserver": null, 
            "range": null, 
            "recursionavailable": null, 
            "redirectportrewrite": null, 
            "redirurl": null, 
            "rhistate": null, 
            "rtspnat": null, 
            "save_config": true, 
            "servicebindings": null, 
            "servicegroupbindings": [
                {
                    "servicegroupname": "test-service-group-1"
                }
            ], 
            "servicetype": "HTTP", 
            "sessionless": null, 
            "skippersistency": null, 
            "sobackupaction": null, 
            "somethod": null, 
            "sopersistence": null, 
            "sopersistencetimeout": null, 
            "sothreshold": null, 
            "ssl_certkey": null, 
            "state": "present", 
            "tcpprofilename": null, 
            "timeout": 10.0, 
            "tosid": null, 
            "v6netmasklen": null, 
            "v6persistmasklen": null, 
            "validate_certs": false, 
            "vipheader": null
        }
    }, 
    "loglines": [
        "Applying actions for state present", 
        "Checking if lb vserver exists", 
        "Add lb vserver", 
        "service_bindings_identical", 
        "Getting configured service bindings", 
        "Getting actual service bindings", 
        "servicegroup_bindings_identical", 
        "Getting configured service group bindings", 
        "Getting actual service group bindings", 
        "len 1", 
        "sync_servicegroup_bindings", 
        "Getting actual service group bindings", 
        "Getting configured service group bindings", 
        "Adding servicegroup binding test-service-group-1", 
        "Checking policy bindings identical", 
        "Getting actual appfw policy bindings", 
        "Getting configured appfw policy bindings", 
        "Enabling lb server", 
        "Sanity checks for state present", 
        "Checking if lb vserver exists", 
        "Checking if configured lb vserver is identical"
    ], 
    "msg": "lb vserver is not configured correctly"
}

The following is the task:

    - name: Create load balancing vserver bound to servicegroup
      delegate_to: localhost
      citrix_adc_lb_vserver:
        nsip: "{{ netscaler_host_ip }}"
        nitro_protocol: "{{ netscaler_nitro_protocol }}"
        nitro_user: "{{ netscaler_nitro_user }}"
        nitro_pass: "{{ netscaler_nitro_password }}"
        validate_certs: "{{ netscaler_validate_certs }}"

        name: "{{ netscaler_virtual_server }}"
        servicetype: HTTP
        ipv46: "{{ vserver_ip }}"
        port: 80
        timeout: 10
        servicegroupbindings:
          - servicegroupname: "{{ netscaler_service_group }}"
      environment:
        no_proxy: "{{ netscaler_host_ip }}"

If I remove the timeout from the task it completes successfully the first time. Something strange with it, but it was repeatable so I thought I would post it so it could get a look. Let me know if you need some additional information.

[giorgos-nikolopoulos]

Hello,

The problem lies with the NITRO API implementation.

It seems that during creation of the lb vserver, when we do a POST request, the value passed for the timeout attribute is ignored and the default value is used instead.

Since there is a check at the end of the module run to verify that all arguments were processed correctly this raises the error you see.

When the playbook is run a second time, we do a PUT request to update the lb vserver attributes. It seems that in this request NITRO applies the value correctly for the timeout attribute. That is why the second run is successful.

The root cause solution would be to fix this in the NITRO API. For now I can suggest the workaround of executing the same task twice, and ignoring the specific error on the first run with:

ignore_errors: true

[andrew80k]

Thanks @giorgos-nikolopoulos. While the use of ignore_errors is an option, I'm reluctant to use it. I have thousands of lines of ansible and might use it once. The problem here is that ignore errors lets the play complete successfully but it's not idempotent. It still makes changes on the subsequent run. I prefer to just remove the timeout for this particular implementation.

[giorgos-nikolopoulos]

Glad I could be of assistance. Cheers.