Jump to content


Photo

TNI v2 requests


  • Please log in to reply
10 replies to this topic

#1 Hemo2

Hemo2

    Advanced Member

  • Members
  • PipPipPip
  • 61 posts

Posted 12 July 2011 - 05:41 AM

I have been getting our techs converted over to TNI v2 and immediately have received the same few comments initially. I have already mentioned these via private messages but I let our folks know I'd post them again.

- I edited all our login scripts and setup the login script scanning with the TNIWINAGENT.EXE program. We already have some users that have logged in many times, creating a new asset file for each login with the 'parentheses' as part of the filename. (Some already have 19 asset files in a few days.) With over 10,000+ computers, you do the math and you can see our concern with the number of files being written to our server that stores the login script scanning files. I have already received questions on how we can control this with only a single inventory file for each computer stored out on our server. Please add an "/overwrite" switch (or some functionality) to the TNIWINAGENT.EXE program so we have the option to only store a single file for each computer regardless of how many times a user has logged in and ran the login script scan.

- What are all the command line switches for the tniwinagent.exe program? (It appears that section of the online documentation is under construction.)

- I know this is already being considered so I apologize for posting again but I told them I'd post this concern. I've received questions on why we are getting duplicate records after a motherboard or computer was replaced or shuffled down, etc. (This is due to the file names getting stored with the 'mac' address as part of the file name.) We use strict naming on our computers and when a computer gets replaced or shuffled down, or a replacement motherboard was put in, we still use the same computer name, but the NIC was obviously changed, which means the MAC address is different, now creating 'different' asset files for that user. Again, with over 10,000+ computers that continuously get replaced, moved, shuffled, repaired, etc., we need some way to NOT store the MAC address as part of the file name and store files the way they worked in v1.6.8 where is was simply the "computer name". (Or store them as COMPUTER_NAME.your.domain.com) That way just because the NIC got changed, the actual inventory file for that particular user only exists once, instead of ending up with seperate hardware inventory files for the same user, which causes problems with reporting and tracking what computers are installed and what software we have installed in offices.

#2 Benny

Benny

    Advanced Member

  • Members
  • PipPipPip
  • 44 posts

Posted 12 July 2011 - 11:35 AM

I know they changed to using a MAC address because they were having problems with duplicates when a computer name changes. When you change out a motherboard, almost everything changes about the computer, even the serial number. If they went back to using the computer's name, they'd have more duplicate issues than using the MAC address. Not sure there's a solution to this. Something has to give, you know?

#3 Hemo2

Hemo2

    Advanced Member

  • Members
  • PipPipPip
  • 61 posts

Posted 12 July 2011 - 12:01 PM

I know they changed to using a MAC address because they were having problems with duplicates when a computer name changes. When you change out a motherboard, almost everything changes about the computer, even the serial number. If they went back to using the computer's name, they'd have more duplicate issues than using the MAC address. Not sure there's a solution to this. Something has to give, you know?


Hi Benny. I thought they changed to using the MAC due to duplicate comptuer names "across domains"? (i.e. - If you have 2 seperate domains, but by chance had a computer named the same in both domains, giving you two completely different computers that actually exist with the same name.) In TNI v1.6.8 you could only store a single one of these records since the file name was simply the 'computer name', thus only one of these could exist in TNI. Using a naming scheme of "computer_name.your.domain.com" would solve both that problem and my problem.

In v1.6.8, when we changed out a motherboard we did not get a 2nd, and duplicate record, since the actual name of the file stayed the same if the MAC addressed changed.

For the purposes of 'asset' management, the current naming scheme using the MAC address does not work at all and all the data in your reports are flawed since you end up with 'old' and 'new' (duplicate) records even though only 1 physical computer exists.

Basically our scenario is we have a computer named HRLA1111. The resulting file name inside of TNI v2 is HRLA1111_1234567890 (or whatever the MAC address is.) We have this asset inventory record now as part of TNI v2. So now, 5 months down the road, the motherboard dies and we replace it with a new motherboard, which changes the NIC/MAC address. We now end up with a "new" asset file inside of TNI v2 of HRLA1111_ABCDEFGH (or whatever the MAC address is.) But keep in mind, the previous record (file) also still exists in TNI.

So we now have 2 physcial records in TNI that show they exist in a particular office, but we have only 1 actual computer. So when you run a report, the data is completely flawed since it doesn't accurately reflect what you actually have. It's going to say we have 2 computers and 2 copies of Microsoft Office, etc., when in fact there is only 1.

A similar scenario is when you buy a new computer to replace an existing one, but 'shuffle down' the existing one to a user. The actual computer name remains the same for a particular user, even though the hardware is different. (Our computers names are setup to reflect the office, city, phone number for the user to make it easy for our techs to quickly identify any computer.) So the new computer comes in and gets the existing name of an old computer. That old computer gets shuffled to a different user, getting the same computer name that person had. Now we have 4 files in TNI, but only 2 physical computers actually exist.

As it is, we have over 10,300 computers to be inventoried and we are constantly replacing, repairing, shuffling, etc. and the naming scheme in TNI v2 is a nightmare for us because we will very quickly end up with 'many' duplicate entries in TNI. (We already have some.) There is no way with so many computers spread across an entire state that we can manually track every hardware change to know that we need to delete the "old" record in TNI v2 when a simple hardware change occured like replacing a NIC. Plus the fact that we have the Nationwide License so we have TNI installed on multiple nodes so you have to try to manage duplicate records on many different computers. (Basically it's not a workable situation.)

#4 Hemo2

Hemo2

    Advanced Member

  • Members
  • PipPipPip
  • 61 posts

Posted 19 July 2011 - 01:26 PM

Wanted to post a quick followup on another reason why we feel having an optional "/overwrite" command line switch available for use with the tniwinagent.exe program.

I have login script scanning setup and it stores the resulting inventory files on a shared server. I have spent the better part of 2 days now just simply 'importing' these files. The import process with so many thousands of files is terribly slow. I assume this is due to the import process opening each and every file across the network and having to look at the 'scan date' of the inventory file to determine if it needs to be imported or not. Since a new inventory file is created each time a user logs into their computer, there are many files per computer that the import process needs to deal with. If we could just have a single file for each computer, I suspect the import process would be much faster.

We could periodically delete all the inventory files on the shared server, to clean it up, but that would be a never ending task and you could never do it efficiently enough to stay ahead of the numerous logins thousands of users do each day. Plus, we have many different techs and managers that have access to these files and we don't all import them at the exact same time, thus we need to leave the inventory files out on this server permanently to ensure they're always available for anyone to import.

I understand why having seperate inventory files created for each login can be useful, but for our situation, a single file getting overwritten with each login scan fits our needs much better.

#5 Hemo2

Hemo2

    Advanced Member

  • Members
  • PipPipPip
  • 61 posts

Posted 28 July 2011 - 05:57 AM

A great big THANK YOU for the "/overwrite" switch in the latest update!

:)

#6 Hemo2

Hemo2

    Advanced Member

  • Members
  • PipPipPip
  • 61 posts

Posted 18 August 2011 - 12:43 PM

Sorry, but I have to revive the issue of duplicate records for the same computer because the file names are stored with the MAC address.

We get duplicate records for machines because of replacing a motherboard (or NIC), we get them when we buy a new computer to replace an existing one and give the new computer the same name as the old one, but now we are getting them on laptops where a user will be physically plugged into the network with the LAN cable plugged into the ethernet jack on the laptop while they are sitting at their desk and then they have to go to a meeting and take their laptop with them and log in wirelessly, which of course the wireless card has a different MAC address which creates a duplicate inventory file for the same computer.

#7 Hemo2

Hemo2

    Advanced Member

  • Members
  • PipPipPip
  • 61 posts

Posted 26 August 2011 - 06:46 AM

I was working on a pc today that had a SQL Data source defined. I looked in the TNI record for this computer and don't think it logs DSN info, so thought I'd add a request to see if it's possible to gather DSN info from the registry in an update.

I believe the list DSN names on a computer are stored in "HKEY_LOCAL_MACHINE\SOFTWARE\ODBC\ODBC.INI\ODBC Data Sources" key. Then for each individual DSN name, the definition/info for each of these is located in a registry key under the "HKEY_LOCAL_MACHINE\SOFTWARE\ODBC\ODBC.INI\" branch and the name of the key matches the name of the DSN retrieved from the "HKEY_LOCAL_MACHINE\SOFTWARE\ODBC\ODBC.INI\ODBC Data Sources" entries. (i.e. - You have a data source named "MyData", so inside the "HKEY_LOCAL_MACHINE\SOFTWARE\ODBC\ODBC.INI\ODBC Data Sources" key, you will find a REG_SZ entry called "MyData" and the definition for the MyData DSN has its' info stored in the "HKEY_LOCAL_MACHINE\SOFTWARE\ODBC\ODBC.INI\MyData" key.

If it's not convenient to retrieve the actual definition info for each DSN, it would still be helpful to retrieve the list of DSN names from the "HKEY_LOCAL_MACHINE\SOFTWARE\ODBC\ODBC.INI\ODBC Data Sources" key. That way if you need to rebuild a computer, you'd at least know they had a DSN defined and you could ask to get the actual definition info from someone so you could recreate that DSN.

#8 Zak

Zak

    Administrator

  • Root Admin
  • PipPipPip
  • 688 posts
  • Gender:Male

Posted 16 September 2011 - 03:30 AM

We will return to the issue of duplicated computers in the nearest days.

Thanks for the idea with SQL data sources, we will try to add them in the future updates.
Softinventive Lab support

#9 Hemo2

Hemo2

    Advanced Member

  • Members
  • PipPipPip
  • 61 posts

Posted 16 September 2011 - 08:34 AM

We will return to the issue of duplicated computers in the nearest days.


Thank you Zak.

#10 PatErauw

PatErauw

    Newbie

  • Members
  • Pip
  • 4 posts

Posted 08 October 2011 - 01:59 AM

Hi,

Maybe the solution to your dupilcate reports due to changed MAC address is fairly easy to solve manually.

I had credential problems remotely scanning a network over VPN, so I run TNI on the clients server, have him scan the network, and than copy the TNI storage files over VPN to my computer at the office to create reports etc. locally.

So I've been experimenting with transporting the TNI storage files between computers.
From what I see there is only a loose connection between the few .xxx files like .folder, .software etc. and the individual *.inv inventory files.

It appears that TNI on startup reads all .inv files to rebuild its storage (correct me if I'm wrong).
Also it appears that when you simply delete an .inv file outside of TNI, the asset (pc) will disappear from the storage and thus from the reports.

This opens up the possibility for you to manually delete the duplicate .inv files jobs from the shared directory where your 10.000+ .inv files are stored, or using some homegrown batch, script or tool that sorts them alphabetically (pcname) and then chronologically, to only retain the most recent files.

Hopefully the development team can shed some light on if this is a correct way to handle the problem and confirm that juggling with the programs file structure does not lead to database corruption etc.

Patrick

#11 Zak

Zak

    Administrator

  • Root Admin
  • PipPipPip
  • 688 posts
  • Gender:Male

Posted 10 October 2011 - 01:38 AM

Hi Patrick,

From what I see there is only a loose connection between the few .xxx files like .folder, .software etc. and the individual *.inv inventory files.

There is a connection between *.inv files and .folder files — the latter being the cache for some primary data from *.inv files. ".software" file holds the software database for the Software accounting.

It appears that TNI on startup reads all .inv files to rebuild its storage (correct me if I'm wrong).
Also it appears that when you simply delete an .inv file outside of TNI, the asset (pc) will disappear from the storage and thus from the reports.

It is so, and it's required to find all new files and remove all records which files were deleted.

This opens up the possibility for you to manually delete the duplicate .inv files jobs from the shared directory where your 10.000+ .inv files are stored, or using some homegrown batch, script or tool that sorts them alphabetically (pcname) and then chronologically, to only retain the most recent files.

That's possible, but it would be more correct to solve this issue using the newly added option to distinguish assets by network name and domain instead of hardware (MAC, etc).
Softinventive Lab support




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users