Server is using a lot more memory since upgrading from LMS 2.6 to 3.2 - any tips?

d-fillmore · ‎08-03-2010

Hi, Can anyone tell me if they have a similar situation to mine

CiscoWorks LMS was upgraded from version 2.6 to 3.2

The Server previously ran at 3.5GB of Memory, but following the upgrade now runs at over 5GB of Memory.

The Server is a HP Proliant DL380 G4.

The Server has very poor response times following the upgrade, and we need to understand why this has occurred.

I guess I need to know if there are some services that you generally disable to improve performance or do we need to upgrade the hardware.

Device has 4G of physical memory with 4G swap size.

Any advice appreciated

Cheers, Dom

Joe Clarke · ‎08-03-2010

LMS 3.2 does take more memory that 2.6. There are more daemon processes which run (e.g. there are now two instances of the DFM servers to support scalability to 5,000 devices). As for how much RAM will be required, that will depend on the number of devices being managed and the other processes running on the server. With 4 GB of RAM, you should be able to manage up to 1.5K devices provided HUM is not installed, and the server is dedicated to LMS alone.

d-fillmore · ‎08-06-2010

Thanks Joseph - It's only managing 152 devices and HUM isn't installed. It was already on support with an SAS contract so I raised it with TAC. We've not got to the bottom of it yet, so far we've doubled the swap size to be 8012Gb.

I'll post the outcome of the case for reference

Cheres, Dom

d-fillmore · ‎09-14-2010

We reduced the size of the log file which was very large and didn't receive anything back from the end user so have assumed that this resolved the issue