SIM 4.2 Data Collection issues

Jeff Cottrill_1 · ‎01-07-2005

I'm still seeing some goofiness in SIM 4.2 data collection, though not near as much as in 4.1

I have a fresh build of 4.2 (not an upgrade from 4.1, and not a migration from CIM7). No DMI or WBEM enabled. The only data collection tasks I have are the default Initial Data Collection and Bi-Weekly Data Collection tasks. No historical data is being collected. Here's what I'm seeing on one server, for example:

Power Supplies:
This server has 2 power supplies. The IM7 DC Report shows 2 Power supplies. The SIM42 report shows 4, 2 listed as "redundant(3)" (same as the IM7 report), and 2 as "nonRedundant(2)"

Installed Controllers:
In the SIM42 report for this server, slots are accurately listed (1 controller, 7 "Empty", and 3 "slot 0" devices). However, there are an additional 8 entries that show board name
"(Empty)", and have no data in the Slot column

None of the above issues is a major probjem. The next one is though...

Logical Disk Drives:
IM7 DC Reports shows the D: drive Total Space at 302537MB, which matches the server's Insight Agents Storage->File System Space Used figures for Total Space
SIM42 DC Report shows the same volume at 288521MB.
The total figure, per Windows, is 317,232,594,994 bytes, which when divided by 1024 twice, gives 302537MB, matching the Agent and IM7 DC report figure.
I checked the R_LogicalDisk view in the SIM42 database and noticed that it gets its MB figure by dividing CIM_LogicalDisk.Win32_Size by 1048576 (1024 twice), which is correct. The CIM_LogicalDisk.Win32_Size value for this volume is 302537000000, which when converted to MB, gives the 288521 figure in the SIM42 DC report. It appears that instead of just leaving that figure at MB, SIM42 is adding zeros to get the bytes figure, rather than multiplying by 1024 twice. In fact, all of the logical disk figures in that table have 6 zeros appended to them instead of the proper byte figures.

Long story short, the disk space figures in the SIM42 reports are not valid and need to be fixed. Would I break anything if I just simply changed the R_LogicalDisk view to divide the figures by 1000000 to compensate for this?

thanks!

Darrin Rawls · ‎01-10-2005

Yep, you found a bug; looks like an update did not make it in. On a running system you can find the substrings in column HandlerInfo of DB table DC_PolledData indexed by the poll ID (19, 20) but you'll have to restart the CMS to make the changes take effect; change the 1000000 to 1048576.

I would not change the view because those variables also hold data from WBEM, DMI, and the NetServer SNMP agents(whose definitions are scaled using 1024), so then the result would be wrong for those cases.

Thanks