ILMT resources?

Hey guys, I am getting ready to install ILMT on one of my relays for audit purposes and I want to make sure I have enough resources on my SQL server where the database will be installed. I was asking my DBA team about it and they asked me the following questions:

“Do you know the size requirements for the new database? Do you know if we would be purging data and how much data would be retained in the database?”

If you could answer any of the above questions, I would appreciate it. I looked at the documentation and I was unable to find this information; however, I did see that if I install on all separate servers, it should be 2GHz, 2 cores for CPU, 8GB memory, and 2BG free HD space. Thank you in advance

The below link may be enough info, didn’t catch this the first time through the documentation :smile:

https://www-01.ibm.com/support/knowledgecenter/SS8JFY_9.2.0/com.ibm.lmt.doc_9.2/com.ibm.license.mgmt.doc/planinconf/r_hardware_requirements_server_windows.html

1 Like

Hey folks, the only piece I do not have is the amount of space the LMT stuff will produce in the SHA1 directory under the UploadManagerData folder. Does anyone have any idea how much space I would need to add 15,000 machines to LMT. Currently, I upload about 2000 machines from SUA each day and it grows about 10MB daily. I am not sure if SUA produces a larger file than LMT. Any experience in this area would be appreciated, I just don’t want to underestimate it and hit a hard limit on the folder setting OR run out of HD space or something. Thank you

For around 20K servers, it takes 1.5 to 2gb of size. So you can adjust accordingly. 2gb looks like a reasonable limit to set.

Why are you installing ILMT on a relay at all? That seems very odd.

I would put ILMT on it’s own VM and not mix it in with a relay.

The majority of the relays in my environment are only relays and have the hardware requirements to meet the ILMT server needs. Also, they are all set up with the proper ACLs on the network. It just sounded easier to do it this way, instead of getting a new VM and starting over and having to do the rest of the setup and work with the other teams on it. Also, when we set up the majority of the environment, SUA, Tivoli RemoteControl, SSP, etc, we installed the majority of these onto relays as well, not all on the same one.

Why would you have it on it’s own box, just for resource purposes? The Database will be external and not on the same relay.

The primary advantage of VMs is that you can have many of them and have them dedicated to a single purpose. If you have dedicated relay VMs with enough resources to run ILMT, then your relay VMs have too many resources and they should probably be lowered.

There are significant security advantages to dedicating a VM to a single task. You can lock it down much tighter for a single purpose. Every purpose you add on to the VM is more holes that need opened.

Also, You can manage resources better. You may need to scale up bulk storage to have a larger relay cache, but you might need to scale up the IOPS of your storage for ILMT in some cases. These are very different storage needs.

The relay might be fine for ILMT now, but in the future, it may need to scale up much more so.

Also, if this relay is hosting the ILMT UI, then the maximum number of users that can connect to that UI will also be affected by the number of clients connecting to the relay due to the number of open TCP connections.

What if you need to reboot the ILMT system many times during troubleshooting or upgrading? This has happened to me many times. What if you then need to reboot the relay that is hosting ILMT due to a relay issue? Having separated VMs minimizes downtime and risk.

What if the relay cache balloons for some reason and causes ILMT to run out of storage? or vis versa?

If you do this enough, then it should be automated and streamlined. If you don’t do it often enough, then do it the right way / harder way. This is never a valid reason to do it a less ideal way, even if it is the reality of your organization.

You should be able to automate the installation and configuration of just about anything through BigFix and/or another combination of tools.


One of the only times I can see putting a relay on something that has other duties is when it is in a location that does not have a datacenter and does not have a way to host VMs. This is typically on the fringes of the network where there aren’t a lot of clients anyway. This is where you might find desktops being used as relays in some cases.

That makes sense, I guess we could have done that with the rest of our stuff as well. I will take that into consideration, thank you for the information.

1 Like

Yep. This applies to basically everything.