Skip navigation

WMI and Remquery data reconciliation

score 35
You have not voted. Not Planned

We have discovered 2 x Windows 2008 Service Pack 1 windows machines with BMC Discovery. We used the "compare to" functionality to view differences between the hosts and noticed that it showed the patches as being different. On host 1 it showed Service Pack 1 as being a patch (unique to this host) and on Host 2 it did not show this patch as being installed. However the OS defined in BMC Discovery for both hosts is Windows 2008 Service Pack 1.

 

We logged on to the hosts and confirmed that they were both Service Pack 1 hosts.

 

When looking at Discovery Access we noticed that on host 1 (which shows service pack 1 as being a patch) the getHostInfo method failed using WMI and obtained all the data (including patches) using REMQUERY. However on host 2 (which does not show service pack 1 as being a patch) the getHostInfo method succeeded using WMI.

 

While Windows Management Instrumentation (WMI) is used as the primary means of discovery; in the event that WMI does not succeed, the proxies use various command line tools via the RemQuery utility. Remquery utility is used as alternate method for the complete discovery.

The fact that we can have inconsistent results in this way can decrease user confidence in the tool.  The tool itself is opting to use either the WMI or the REMQUERY for whatever reason - as a consumer, this is irrelevant to me and I expect the Discovery tool to reconcile and report the correct findings in a consistent manner.  I would be of the view that since the BMC Discovery solution uses both WMI and Remquery to discover the required host information then if the two methods are producing different / conflicting outputs for the same patchset (i.e. SP1) in this case, then this is an issue for the BMC Discovery solution to reconcile and produce consistent findings.  If as a consumer we cannot have confidence in the output of BMC Discovery then it offers diminished value.  Some form of reconciliation should take place and if the solution cannot perform a reconciliation of data then it should either 1) not allow a comparison of two hosts where the information was gathered using different methods; or b) at a minimum inform the user that the information was attained using different methods and there may be inconsistencies.  The point re the latter again comes back to a consumer not being able to have confidence in the findings.

 

Please note that case # 00196891 relates to this issue.

Comments

Vote history