ComputeSensor not sending information to Cloud Optimizer

Hello experts,

We have a CO 3.01 deployed and we want to monitor some Windows hosts. After installing the ComputeSensor as indicated in the guide we found we have no data in CO. We checked the port 381 on both sides but it seems to be OK (picture attached).

One thing we found when enabling the CS service is that this error appears, despite having "regBBC=false"

3: INF: Tue Feb 07 18:59:40 2017: (19940): g_BbcClient is NULL!!

Please find attached the hpcs.conf file of the windows server.

We have opened the port in the firewall and seems to have connectivity. Any suggestion on what else to check? Anything else to configure in host or CO server?

Thanks in advance,

Angel

Parents
  • Hello Angel,

    It is unclear if you are working with Windows Hosts that are part of a VMware environment (VMs) or if you are working with Physical Hosts.

    The ComputeSensor utility is specifically for VMs and intended to provide drill-down guest information for VMs within a VMware environment.  If these Windows hosts are within a VMware environment, then the vCenter server would need to be added as a Datasource to the CO management server.

    If these Windows hosts are Physical Servers, then there is a Physical Server collector that you would install on these systems.  You would go to the CO Management Server > Settings > Add Datasource > Physical Server to see the 'Collector' links for Physical Servers.  You will need to remove the ComputeSensor utility from the Windows Servers before installing the Physical Server Collector.

    More information about the process is available within the CO 3.01 Online Help Guide (P. 37).
    https://softwaresupport.hpe.com/km/KM02469451/Cloud_Optimizer_3.01_OnlineHelp.pdf

    I hope that this helps clarify your situation.

    Please let us know if you have any other questions.

    Regards,

    Mark

  • Hello Mark,

    Thanks for the explanation, trully aprecciate that since I'm just beginning with CO.

    We are working with physical hosts (Microsoft Windows Server 2012 R2 Datacenter) and yes, at first I tried executing the Compute Sensor installer, assuming it was the same thing. Now I have installed the Collector following the instructions for the physical server (screenshot attached). Actually the Windows Service created is the same in both cases, VM and physical, so it is difficult to see the differences.

    Unfortunately, there is nothing regarding this server on CO (in Treemap for example) One thing I found after executing the service in debug mode is that the problem may be that no data is being sent:

    4: DBG: Tue Feb 07 23:53:53 2017: (5484): CSubscriberMap::PostData: Data has been sent for 1 elements
    4: DBG: Tue Feb 07 23:53:53 2017: (5484): URI: 10.232.9.111:8081/.../HPCSListener -> Flush response
    4: DBG: Tue Feb 07 23:53:53 2017: (5484): URI : 10.232.9.111:8081/.../HPCSListener -> Header response HTTP/1.1 100 Continue

    HTTP/1.1 200 OK
    Server: Apache-Coyote/1.1
    Content-Length: 0
    Date: Tue, 07 Feb 2017 22:53:52 GMT

    But accessing to http://locahost:381 there are the metrics as usual.

    Is that Content-Length normal? Is there any place where I can investigate that?

    Thank you very much for your help.
    Angel.

  • Hello Angel,

    Thank you for the confirmation of your configuration.

    Even though the services have the same name, they are not the same.  That is why the documentation states that having the ComputeSensor and Physical Collector on the same system is not supported.  That is also why I asked that you remove teh ComputeSensor software before installing the Physical Collector.

    Please also note that the port being used to transmit data is not port 381.  Port 8081 is used to collect data for Cloud Optimizer.  You may want to confirm that this port is open.

    Using port 381 does confirm that the Physical Server Collector is installed successfully.

    There is also a log file on the CO management server that would help confirm if the data is being received.
    /var/opt/OV/log/status.PhysicalServer

    Since port 381 is functioning properly, the Physical Server Collector is installed and working.  If there is no real data within the CO management server and the status.PhysicalServer log file, then I suspect that there is a firewall issue between the Windows Hosts and the CO management server.

    I just configured the firewall within my setup and it then allowed the communication and the Physical Servers showed up.

    Within the C:\Program Files\HP\HP BTO Software\hpcs\hpcs.conf file, there is a section for hpcs.subscribers.  At the end of this section, there should be an entry similar to the following:

    http://<IP Address of CO Server>:8081/syspulse/HPCSListener=/Metrics/All

    Once I configured port 8081 within the Outbound rules of the Windows firewall, it worked for me.

    I hope that this is helpful.

    Regards,

    Mark

  • Hello Mark,

    Thanks for your answers. I have checked but both servers are in the same network so the firewall is not blocking anything. Looking at the log you mentioned, it shows it has not recognized the servers as it appears "Number of Physical Hosts = 0".

    2017-02-09 15:04:25,973:INFO:Number of Physical hosts 0

    2017-02-09 15:04:25,984:INFO:Execution of syspulse data logging completed for current interval.

    2017-02-09 15:09:25,737:INFO:Starting collection for PhysicalServers

    2017-02-09 15:09:25,737:INFO:Verifying is Listener running

    2017-02-09 15:09:25,745:INFO:Listener is running.

    2017-02-09 15:09:25,745:INFO:Uploading data into stage table

    2017-02-09 15:09:25,952:INFO:Upload utility output is.....  Importing Performance Data from Physicak Systems into Vertica Database

    vsql:/var/opt/perf/syspulse/scripts/importdata_global.sql:4: INFO 2372:  Cannot commit; no transaction in progress

    Class : global  - 30

    30 Records - SUCCESS

    vsql:/var/opt/perf/syspulse/scripts/importdata_disk.sql:3: INFO 2372:  Cannot commit; no transaction in progress

    Class : disk  - 360 Records - SUCCESS

    Class : cpu - 0 Records - NODATA_AVAILABLE

    vsql:/var/opt/perf/syspulse/scripts/importdata_netif.sql:3: INFO 2372:  Cannot commit; no transaction in progress

    Class : netif  - 240 Records - SUCCESS

    vsql:/var/opt/perf/syspulse/scripts/importdata_filesystem.sql:3: INFO 2372:  Cannot commit; no transaction in progress

    Class : filesystem  - 30 Records - SUCCESS

     

    The log repeats these entries all along the file.

    Any other suggestion on where to look?

    Thank you very much for your time.

    Regards,

    Angel

Reply
  • Hello Mark,

    Thanks for your answers. I have checked but both servers are in the same network so the firewall is not blocking anything. Looking at the log you mentioned, it shows it has not recognized the servers as it appears "Number of Physical Hosts = 0".

    2017-02-09 15:04:25,973:INFO:Number of Physical hosts 0

    2017-02-09 15:04:25,984:INFO:Execution of syspulse data logging completed for current interval.

    2017-02-09 15:09:25,737:INFO:Starting collection for PhysicalServers

    2017-02-09 15:09:25,737:INFO:Verifying is Listener running

    2017-02-09 15:09:25,745:INFO:Listener is running.

    2017-02-09 15:09:25,745:INFO:Uploading data into stage table

    2017-02-09 15:09:25,952:INFO:Upload utility output is.....  Importing Performance Data from Physicak Systems into Vertica Database

    vsql:/var/opt/perf/syspulse/scripts/importdata_global.sql:4: INFO 2372:  Cannot commit; no transaction in progress

    Class : global  - 30

    30 Records - SUCCESS

    vsql:/var/opt/perf/syspulse/scripts/importdata_disk.sql:3: INFO 2372:  Cannot commit; no transaction in progress

    Class : disk  - 360 Records - SUCCESS

    Class : cpu - 0 Records - NODATA_AVAILABLE

    vsql:/var/opt/perf/syspulse/scripts/importdata_netif.sql:3: INFO 2372:  Cannot commit; no transaction in progress

    Class : netif  - 240 Records - SUCCESS

    vsql:/var/opt/perf/syspulse/scripts/importdata_filesystem.sql:3: INFO 2372:  Cannot commit; no transaction in progress

    Class : filesystem  - 30 Records - SUCCESS

     

    The log repeats these entries all along the file.

    Any other suggestion on where to look?

    Thank you very much for your time.

    Regards,

    Angel

Children
  • Hello Angel,

    It's good to know that the two servers are on the same network, so the network firewall is not causing a problem.  However, there is still the Windows Firewall that is part of the Windows OS.

    When I looked at my environment, I saw the same messages about the number of Physical Hosts being 0.  It wasn't until I created a new Outgoing rule within my Windows Firewall for port 8081 that this log file entry changed on the CO management server.

    Basically, what is occurring is that the information is not reaching the CO management server from the Physical Host.  You have confirmed that the Physical Host Collector is running by accessing it with the local browser on port 381.  You can also see that the information is not reaching the CO management server based on the status.PhysicalServer.  This communication uses port 8081 between the two servers.

    Please let us all know the status of the local Windows Firewall and the results once port 8081 is opened within the Outgoing Rules.

    Thanks,

    Mark

  • Hi Mark,

    Unfortunately, Windows Firewall is disabled in the server.... (screenshot attached)

    Any idea? Any logs to check maybe in the monitored server?

    Thank you very much.

    Best regards,

    Angel

  • Hello Angel,

    I apologize for the confusion caused by my last entry.

    On the Physical Windows Host where you have the Physical Collector installed, there are log files and entries.  I thought I mentioned these before.

    The Physical Collector is installed into the following location:
    C:\Program Files\HP\HP BTO Software\hpcs

    Within this directory, there are two important files to review.

    • hpcs.conf
    • hpcstrace.log

    You can add debugging to this file with the following setting:
    [hpcs.trace]
    ; Debug_Level=INFO

    However, as I look more closely at the messages you provided from the status.PhysicalServer file, it looks like there is something else going on also.

    It looks like information is being found and that it is having problems with the transactions.  I see some of the same sql errors within my enviroment, but then there is this section in the middle of the errors:
    ...
    2017-02-07 17:00:50,847:INFO:Number of Physical hosts 14
    2017-02-07 17:00:51,038:INFO:PhysicalServers target has been added.
    2017-02-07 17:00:51,040:INFO:Starting observation list creation for data submission
    2017-02-07 17:00:51,041:INFO:Uploading observation list created
    2017-02-07 17:00:52,381:INFO:Successfully sent json data upload request.
    2017-02-07 17:00:52,381:INFO:Execution of syspulse data logging completed for current interval.
    2017-02-07 17:05:50,317:INFO:Starting collection for PhysicalServers
    2017-02-07 17:05:50,318:INFO:Verifying is Listener running
    2017-02-07 17:05:50,330:INFO:Listener is running.
    2017-02-07 17:05:50,331:INFO:Uploading data into stage table
    ...

    Then the sequence ends with the following entries:
    ...
    2017-02-07 17:05:50,828:INFO:Number of Physical hosts 30
    2017-02-07 17:05:50,839:INFO:Starting observation list creation for data submission
    2017-02-07 17:05:50,841:INFO:Uploading observation list created
    2017-02-07 17:05:52,794:INFO:Successfully sent json data upload request.
    2017-02-07 17:05:52,794:INFO:Execution of syspulse data logging completed for current interval.

    If you continue to have problems, then I would suggest opening a support case.

    You would be asked for the log files from the Physical WIndows Hosts as well as the data from the CO management server.  Here's a utility that you can use to collect the necessary information from the CO management server.

    Cloud Optimizer Troubleshooting Toolkit (PVTK) description and FAQhttps://softwaresupport.hpe.com/km/KM02600579

    I hope that this helps.

    Regards,

    Mark

  • Hi Mark,

    Thanks for your indications, I opened a support case to continue investigating this issue in more detail. I will post the advances and if we eventually solve this.

    Again thank you very much!

    Best regards,

    Angel