ZABBIX BUGS AND ISSUES
  1. ZABBIX BUGS AND ISSUES
  2. ZBX-10257

VMWare item would not get NOTSUPPORTED if VMWare collector was not started

    Details

      Description

      When setting up VMWare monitoring if forgot to add StartVMwareCollectors option to Server/Proxy configuration file an item would not become NOTSUPPORTED. E. g.

      • create vmware item e. g. vmware.fullname[https://192.168.10.10/sdk]
      • make sure StartVMwareCollectors is not specified in Server/Proxy configuration file
      • start Server/Proxy

      An item will not receive any data and will not get NOTSUPPORTED state. Expected result would be a NOTSUPPORTED state item with error message something like VMWare collector not started.

        Activity

        Hide
        Viktors Tjarve added a comment - - edited

        Fixed in development branch svn://svn.zabbix.com/branches/dev/ZBX-10257

        Show
        Viktors Tjarve added a comment - - edited Fixed in development branch svn://svn.zabbix.com/branches/dev/ZBX-10257
        Hide
        Andris Zeila added a comment - - edited

        (1) It would be better to check if CONFIG_VMWARE_FORKS is not 0, rather than using zbx_vmware_get_statistics(). This way we would avoid of unnecessary vmware cache locking if collector(s) are running.

        We should use lowercase letters if we are talking about our vmware collector. All of our processes have lowercase names - so we kept similar naming style for vmware collector. If we are talking about VMware as a product - we should use their capitalization (VM - uppercase, ware - lowercase) - VMware (for example VMware checks).

        In this case the message is about our collector process - so it should be something like 'vmware collector is not started', or '... not running'. We have messages 'no "process type" processes started', but those messages are used for checking the specific processes - the situation is not quite the same. Maybe we need to double check it with Martins what would be the best choice.

        The error messages that are show in frontend should start with uppercase letter and end with .. It would be nice to fix also the rest of error messages in checks_simple.c file.

        Viktors Tjarve RESOLVED. Please review the r58010 if you agree with using get_process_type_forks() for process check rather then just checking if CONFIG_VMWARE_FORKS is not 0.

        Andris Zeila CLOSED

        Show
        Andris Zeila added a comment - - edited (1) It would be better to check if CONFIG_VMWARE_FORKS is not 0, rather than using zbx_vmware_get_statistics(). This way we would avoid of unnecessary vmware cache locking if collector(s) are running. We should use lowercase letters if we are talking about our vmware collector. All of our processes have lowercase names - so we kept similar naming style for vmware collector. If we are talking about VMware as a product - we should use their capitalization (VM - uppercase, ware - lowercase) - VMware (for example VMware checks). In this case the message is about our collector process - so it should be something like 'vmware collector is not started', or '... not running'. We have messages 'no "process type" processes started', but those messages are used for checking the specific processes - the situation is not quite the same. Maybe we need to double check it with Martins what would be the best choice. The error messages that are show in frontend should start with uppercase letter and end with . . It would be nice to fix also the rest of error messages in checks_simple.c file. Viktors Tjarve RESOLVED. Please review the r58010 if you agree with using get_process_type_forks() for process check rather then just checking if CONFIG_VMWARE_FORKS is not 0. Andris Zeila CLOSED
        Hide
        Andris Zeila added a comment - - edited

        Successfully tested, please review r58073 where after some discussion was decided to use the same error message as for internal checks.

        Viktors Tjarve Looks good.

        Show
        Andris Zeila added a comment - - edited Successfully tested, please review r58073 where after some discussion was decided to use the same error message as for internal checks. Viktors Tjarve Looks good.
        Hide
        Sandis Neilands (Inactive) added a comment - - edited

        Released in:

        • pre-2.2.12rc1 r59008;
        • pre-3.0.2.rc1 r59009;
        • pre-3.1.0 (trunk) r59010.
        Show
        Sandis Neilands (Inactive) added a comment - - edited Released in: pre-2.2.12rc1 r59008; pre-3.0.2.rc1 r59009; pre-3.1.0 (trunk) r59010.
        Hide
        Sandis Neilands (Inactive) added a comment - - edited

        Documented in https://www.zabbix.com/documentation/2.2/manual/introduction/whatsnew2212 . Martins Valkovskis, please review and add it to 3.0.2 "What's new" page when it's created.

        Martins Valkovskis Reviewed, with small grammar changes. Added also to: https://www.zabbix.com/documentation/3.0/manual/introduction/whatsnew302#daemon_improvements. RESOLVED

        Sandis Neilands CLOSED.

        Show
        Sandis Neilands (Inactive) added a comment - - edited Documented in https://www.zabbix.com/documentation/2.2/manual/introduction/whatsnew2212 . Martins Valkovskis , please review and add it to 3.0.2 "What's new" page when it's created. Martins Valkovskis Reviewed, with small grammar changes. Added also to: https://www.zabbix.com/documentation/3.0/manual/introduction/whatsnew302#daemon_improvements . RESOLVED Sandis Neilands CLOSED.

          People

          • Assignee:
            Unassigned
            Reporter:
            dimir
          • Votes:
            0 Vote for this issue
            Watchers:
            6 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved: