We are using CiscoWorks LMS 4.0.1.
Acquisition info reports last acquisition end time 19 July and Campus data collection running.
Number of host entries is 68 but we have two thousands PCs.
When we try to perform an acquisition action with scope all hosts and users, we receive this message
Failed to start acquisition: Construction of XML data required for UT is in progress.Please try after some time.
In the 'collection summary' do you still see the processes for data collection and UT acquisition complete from time to time?
If data collection keeps on running then UT work start.
Hello Michel and many thanks for your help.
I apply the command to reinitialize the CM database and now user acquisition works fine.
after some acquisitions I can verify that many many entries are coming from the same device.
No Andrea, just questions :-)
In the 'collection summary' do you see if UT is collecting from all devices?
Same number of devices as data collection and other modules?
Are some of the devices on which you don't get endhosts Nexus machines?
Are you using SNMP V2c or V3?
OK, so now you see many devices on the backbone switches I presume?
I know you can configure UT to discover endhost on trunkports but I leave that off normally. You may want to check that.
Also if there is a @ in your community strings you get an issue with UT
This is because of the way it collects bridgetables for vlans using community@vlanID as a community string.
Good evening Michel.
You are right. I can see many hosts on backbone but acquisition on trunks is disabled and there is not special character on community string.
Sent from Cisco Technical Support iPad App
So when you run a UT report for a switch the report is completely empty?
If so that would indicate LMS was unable to get the mac tables of these switches.
I have not tried to much with this myself but the UT logs should indicate a problem if it can't get any info from the device, if not, increasing the loglevel to debug would.
But If the UT would fail it should show this in the collection summary, but there everything seems OK. Strange.
The onlt other thing that comes to mind is that the switches forget too soon the mac addresses that LMS tries to get.
This could hapen if frequent root topology change would occur. This put the 'forget' timer on the switches really low.
Maybe check the last time the switch saw a topology change in the CLI.
Not too much other ideas what else could cause this.
I have verified UT acquisition settings without success.
UT continues to report many hosts on backbone switches.
End-host remaining on trunks can be caused because the switch to which they are connected is not managed or does not support UT.
Some nexus don't support UT (yet) I think
You should try to locate some of the end-hosts your self.
Is the switch where they are connected managed by LMS, not one of the 3 failed ones? Does that switch support UT.
You can find this information in the supported devices table.
You went frol 68 end-hosts to 3215 end-hosts. How many do you expect?