DX Infrastructure Management

Expand all | Collapse all

Qos Discovery not happening properly

  • 1.  Qos Discovery not happening properly

    Posted 08-11-2017 10:54 AM

    Hi All,

     

     We have discovered a server with robot and as well as vmware too, but metrics is not coming for robot metrics in UMP home.

     

    Niscache and robot id reset has been done .

    Server has removed using cs_key too.

     

    Kindly share that anything needs to checked 



  • 2.  Re: Qos Discovery not happening properly

    Posted 08-11-2017 11:36 AM

    Hi Issac08

     

    The above steps you've taken are correct however, we would also null the ci_metric_id and restart the discovery_server & data_engine probes - have you also done that?. If not, I would recommend trying the following for this device and if the problem still persists raise a support ticket with the requested information attached for us to investigate further. 

     

    1. Run the following for the device that does not display data (note the cs_key value as it will be used in the next step)


    SELECT name,ip,origin,cs_key FROM CM_COMPUTER_SYSTEM WHERE ip = '<ip>';

     

     

    2. Remove the device. Select the discovery_server probe and press Ctrl-P from your keyboard. On the probe command set select 'remove_master_devices_by_cskeys'. On the bottom of the probe utility GUI you will see the csKeys input box, place the affected device "cs_key". Press the green play button on the probe utility and it will execute the command.


    3. Clear robot niscache. Within IM, navigate to the controller probe on the robot in question. Open up the probe utility. Open the Options dialog by pressing the second to the right icon. In options, select the "Expert Mode" checkbox and press OK. In the probe commandset, select "_nis_cache_clean" and press the green arrow button to run it.


    4. Again, on the controller probe for the robot in question. Open the probe utility. In the probe commandset, select "_reset_device_id_and_restart" and run it.


    5. In the database run the following: update s_qos_data set ci_metric_id = NULL where source = '<source_ip_address>'


    6. On the primary hub increase discovery_server log level to 3 (also increase log-size), restart the discovery_server and data_engine probes.


    Once all of the above is complete, go back into USM and check if the server displays any metrics. If not, please raise a support ticket and attach the discovery_server.log and _discovery_server.log. This should provide information regarding the correlation (Provide hostname and IP so we can use this when searching the logs). 

     

    Hope this helps. 

     

    Kind regards, 

    Ryan 



  • 3.  Re: Qos Discovery not happening properly

    Posted 08-11-2017 01:58 PM

    It might be helpful to first go to Performance Reports Designer > select the robot > and then see if any qos and metrics show up for it.

     

    For data to show up in USM associated to a device it needs things like dev_id & met_id, but for some reason PRD can show the data even without. so it's a quick easy test to see if data is collected.

     

    Of course drnimbus can also be used, message sniffer filtered to the robot.

     

    If you want to check the db directly

    to see if there is qos

    select * FROM [CA_UIM].[dbo].[S_QOS_DATA] where source = 'robot name'

    from that you get the table_id & rn table and plug those in

    as an example

    select * FROM [CA_UIM].[dbo].[RN_QOS_DATA_0139] where table_id = 2745

    and that would show you the data points



  • 4.  Re: Qos Discovery not happening properly

    Posted 08-28-2017 07:27 AM

    Qos is populating from robot and able to see data in Performance Reports Designer ,but not able to see data in Home->Unified Service Manager .Same time i could see see two entries in inventory one from vmware probr and other from robot correlation is not happening here.



  • 5.  Re: Qos Discovery not happening properly

    Posted 08-29-2017 07:25 AM

    If you are using UIM 8.4 or greater, set the following key in data_engine.cfg

     

    <setup>/update_metric_id = yes

     

    Metric_id mismatch is expected to be solved automatically with this key.

     

    Regards,

    Yu Ishitani



  • 6.  Re: Qos Discovery not happening properly

    Posted 08-29-2017 10:23 AM

    Thanks ,it has worked .But why this entry is required ?



  • 7.  Re: Qos Discovery not happening properly

    Posted 08-29-2017 10:33 AM

    The update_metric_id key instructs the data_engine to update the metric_id when it is changed. By default, the key is set to no, or not present in the configuration. See the following: Correlation Issues with Metric Data - https://docops.ca.com/ca-unified-infrastructure-management-probes/ga/en/alphabetical-probe-articles/data_engine/data_engine-troubleshooting#data_engineTroubleshooting-CorrelationIssueswithMetricData 

     

    Kind regards, 
    Ryan 



  • 8.  Re: Qos Discovery not happening properly

    Posted 08-29-2017 10:36 AM

    Thanks.



  • 9.  Re: Qos Discovery not happening properly

    Posted 08-29-2017 08:52 PM

    I apologize I did not describe the reason of asking.

     

    There are multiple cause for the symptom you described.

    One famous cause is "Mismatch of metric_id".

    In USM we are running JOIN sql statements in the background and sql will return nothing for list of metrics in case Mismatch of metric_id.

     

    The key in data_engine.cfg is helpful to automatically fix the situation of Mismatch of metric_id.

     

    Regards,

    Yu Ishitani