CC3 --> Vanilla Windows (without wiping?)

[CyBeRkId2002]

Hi all,

 

In the coming months I have the job of planning and executing a move from CC3 to vanilla windows (cost saving exercise )

 

I have been over this a thousand times in my head but I am still not sure on the best way to proceed. As a bit of background we have quite a complex network and am reluctant to start from scratch (the main area's I am worried about - mainly due to lack of knowledge and understanding - are connecting drives back up to the SAN and a few other bits and pieces).

 

As such I am now thinking about doing the following:

 

• Removing all RM/CC3 components from our 3 RM servers (we have other non-RM servers)
  
• Installing RIS to handle client builds
  
• Installing/configuring WSUS to handle updates
  
• Creating a new 'master' OU with inheritance disabled which will serve as the base for our new policies and GPOs. Sort out and test package allocation from here.
  
• Sort system for allocating drives and printers
  

 

What are peoples thoughts on my (very basic) plan above. As a side-note we have no performance issues at all so I am not too concerned about leaving a few old RM components around as long as they do not interfere with the running of the network.

 

Thanks in advance for any advice and feedback,

 

Michael

[irsprint84]

I got for FOG with client builds

[bart21]

+1 for fog

[Hightower]

Wouldn't even try, you will probably encounter an array of problems. An ex-colleague tried to remove all the CC3 components from a laptop to make it a Vanilla client here, and it ran dog slow with various problems - so really wouldn't suggest doing that to live servers.

 

Plan, plan, and then plan. Once you've planned it, plan some more and then take a holiday. When you come back, plan some more and then flatten the network and start from scratch. It really will be a whole world easier, and you'll have a stable network from it (not saying RM is not stable, saying that removing the RM stuff from the servers could make it unstable).

[webman]

How are you going to save money moving away from an existing CC3 network, where all licenses are still valid? Your investment in CC licenses will have been wasted. Do you use RM support? If not, don't pay for it. Just think of all the third-party software (and your time!) needed to replicate everything you already have included with CC3.

 

If you do insist on moving, I'd definitely recommend what Hightower says - don't bother uninstalling - more hassle than it's worth.

[meastaugh1]

Michael, I am pursuing a similar CC3 exit path to the one you suggest. The majority of our desktops are now running vanilla XP with WSUS for updates and WDS for OS deployment. We are just about to start pushing out Win7 with MDT2010. We have setup new OUs at the root and applied GPOs with loopback enabled, allowing us to slowly transition away from C3.

 

I'm still to finalise a plan for the servers but my current thinking is to add some 2008r2 DCs and transfer across FSMO, DHCP, DNS etc then demote the old RM DCs to member file servers, then remove/disable CC3 server components.

[CyBeRkId2002]

Hi all,

 

thanks for the responses!

 

@Hightower: thanks for the reply. Have been looking into the one area which worried me the most today (connecting drives to the SAN via iSCSI and doesn't seem as bad as expected). Will definately give flattening more consideration.

 

@webman: The cost saving is in the support. Every now and again we come up against an issue related to the RM software which we cannot solve without logging a support call. What do we do then without a support contract? Also, certain windows updates can cause issues with the RM network and as such, rather than using WSUS they release update patches every month or so with 'approved' MS hotfixes. Without support you are not entitled to these and would require manually checking, downloading and allocating every month.

 

@meastaugh1: How are you finding your network performance wise? Did you have any trouble running and configuring WDS and WSUS alongside CC3? I suppose my main concern is running into things which RM have locked down on the server but I imagine these are mainly within the Domain Controller GPO. Any other tips you can provide as you seem to be going through the exact process we are looking at.

 

Thanks all for your time replying,

Michael

[jamiesev]

Hi All,

 

I am also in the process of removing CC3 from our network. I have made up new OU's with inheritance disabled and setup associated GPs for them. I then ested GPs on both RM and vanilla desktops. All seem ok. The next move for us is to reimage all the computers back to vanilla and then decomission and rebuild our servers one at a time. It would be good to keep contact between us so we can let each other know of any issues we may have run into! Please feel free to PM me for my email address.

 

Good luck!

 

Jamie

[meastaugh1]

@CyBeRkId2002 on the whole it's fine. The only significant issue is that it can take a few more seconds than I'd like when a user logs on to a computer for the first time. This is because there seemed to be problems with logging on to a vanilla workstation with a CC3 profile, so until migration of the desktops is complete, I've set the vanilla desktops to use local profiles. No CC3 specific issues with WSUS or WDS that I can recall, as long as you've got them installed on separate (physical or virtual) servers. The only issue I've come up against, where the the RM DC GPO is the cause, is when I setup a vanilla terminal server and placed the server in the Servers - No Inheritance OU, as this still has the RM DC GPO linked to it. This caused a problem because this policy sets permissions on explorer.exe to only allow adminitrators access to this file (to prevent non-admins logging on to servers). This was fixed by changing the GPO and would only affect you if you're setting up a TS.

 

Let me know if you've got any more specific questions and I'll try to help/advise.

[Hightower]

To me, this honestly sounds like a headache. Someone having to use local profiles? What's that about!

 

Seriously, you have two options. Stick with CC3 or go vanilla with a complete new install. There is no half way house of removing the RM stuff in my eyes - that is just begging for a call from Mr Argghhh Why The Frick Did I Do That?!

[TechMonkey]

Same as Hightower, I can't see any benefit to running a 'dirty' system if you have the chance to restart. You will constantly be chasing bugs or phantoms caused by some remnant of CC3 & even if it is specifically that you won't be able to be sure.

[Norphy]

How are you going to save money moving away from an existing CC3 network, where all licenses are still valid? Your investment in CC licenses will have been wasted. Do you use RM support? If not, don't pay for it. Just think of all the third-party software (and your time!) needed to replicate everything you already have included with CC3.

 

If you do insist on moving, I'd definitely recommend what Hightower says - don't bother uninstalling - more hassle than it's worth.

 

Speaking from our experience, we're running CC3 at the moment and will be moving to a Windows 7/2008 R2 system next summer. We won't be going for CC4. Why? It'd cost us the better part of £70,000 to buy the licences and CALs for CC4 alone. We can get replacements for all of the CC3 functionality that we use for barely a third of that plus frankly the tools we're getting do a much better job than RM's. Yes, moving from CC3 will take a lot of time, effort and planning but considering that Microsoft practically give you tools like SCCM and that we're already using tools like Papercut for printer management, why exactly would we want to pay through the nose for CC4? Especially in these austere times?

 

The other advantage is that we can set up our network in the way that we want it to be set up rather than in the proscribed CC3/4 method. Don't get me wrong, I think CC is a good product for those who need it but for us at least, it's a square peg trying to be forced into a round hole. As my colleague @Roberto says, that's not the fault of the hole or the peg but either way it isn't a good fit.

 

I think blatting the entire network is a tad excessive. We're going to be keeping our domain in place but the domain controllers and workstations will be rebuilt or replaced and no trace of the CC3 software will remain on either. Member servers are already non-CC3.

[spc-rocket]

Yes i agree with the idea of doing a clean install and migrating the users and resources over. It is a tall order if you have lots of complexity with the CC3 network but should be worthwhile when its working correctly. Having said this not everyone has the time to plan it out or sometime the resources to do this (i'm not going to mention the skillset because i feel that schols nowadays have really good NMs and Technicians that can manage vanilla networks as well as other such as RM , Ranger, CSE etc).

 

There are tools out there that replaces some of the functionality the RM system provides. The hiding of shortcuts when not installed on the PCs, reset of profiles and permissions can be done quickly now with various tools or scripts so its not a problem. Norphy also mentioned SCCM (the CAL - this is the thing that costs) is included in the EES Enterprise pack so it reduces the costs even more because EES is licensed on the FTE count of staff rather than the number of station a school may have.

 

A new domain might lok like a complex idea but with trust relationship setup iniatially to migrate the groups and users over it should be okay for most of the part. I believe RM have now also started offering support for vanilla networks as well so they know certain amount of schools will be leaving RM and not moving to CC4. This is fair enought because the support of RM is really good we found them to be very good when we had issues.

 

 

 

 

Speaking from our experience, we're running CC3 at the moment and will be moving to a Windows 7/2008 R2 system next summer. We won't be going for CC4. Why? It'd cost us the better part of £70,000 to buy the licences and CALs for CC4 alone. We can get replacements for all of the CC3 functionality that we use for barely a third of that plus frankly the tools we're getting do a much better job than RM's. Yes, moving from CC3 will take a lot of time, effort and planning but considering that Microsoft practically give you tools like SCCM and that we're already using tools like Papercut for printer management, why exactly would we want to pay through the nose for CC4? Especially in these austere times?

 

The other advantage is that we can set up our network in the way that we want it to be set up rather than in the proscribed CC3/4 method. Don't get me wrong, I think CC is a good product for those who need it but for us at least, it's a square peg trying to be forced into a round hole. As my colleague @Roberto says, that's not the fault of the hole or the peg but either way it isn't a good fit.

 

I think blatting the entire network is a tad excessive. We're going to be keeping our domain in place but the domain controllers and workstations will be rebuilt or replaced and no trace of the CC3 software will remain on either. Member servers are already non-CC3.

[irsprint84]

When you have both vanilla and CC4 up to the spec you want they can equally run as well as each other but with CC4 its crazy costs! I m sure RM are trying to price themselves out, 6-8 years ago a RM school wouldnt dream of moving away from RM now I think the cost is driving people away

[Hightower]

It doesn't take much to turn a simple question (how to go vanilla without wiping) into an RM v Vanilla thread. If people want to contribute to those topics why not find a thread about that, and stop turning these threads into the same discussion (is it about once a week these discussions come up?!)

[irsprint84]

Do fresh install I did that with CSE I put it simply, would you rather do a fresh Windows 7 install or use a manufacters build?

[Roberto]

I think it is perfectly possible to do a CC3-->vanilla move within the current domain, and its also a perfectly valid model to migrate to a clean domain too but either way it depends how comfortable you are with AD. Either way you will be looking to either do a lot of hard work on a migration or a lot of hard work untangling the current domain. Something that might make your decision easier is thinking about the other services in your domain that rely upon domain services and the difficulty of migrating those to a new domain (and whether or not it is desirable to migrate them anyway, e.g. as part of an upgrade to these systems alongside your main network upgrade). Oh and think about how confident you are in the current health of your domain/forest as they stand now. If you think it is robust then that might suggest staying with what you have, and if you think it is problematic this might favour a migration.

 

As @Norphy says, we've chosen to stay with our current domain/forest (we actually have a 2 domain forest with CC3 in the parent domain). I'd suggest there's a cost no matter what you do and its just a case of choosing where you are most comfortable with this cost hitting you.

 

To stay within the current domain you will need to create a new OU structure and, I would suggest, a complete new set of GPOs. You'll obviously need to replace the GPOs that were in place for CC3, that goes without saying, but even for your own custom ones I would suggest starting again with fresh ones because the old ones will have been designed based on assumptions about your current environment. You then need to consider what will need to change with regards to user account settings (e.g. profile, etc) when you move them to the new OU structure and you'll need to create a 'test' network to test and practice the migration process.

 

I'd also suggest that you must rebuild all the DCs rather than simply attempt to uninstall things. This is the easiest way to clean up the CC3 software/services that will be running on them - you can create new domain controllers to take over the DC functionality from the current CC3 DCs when the time is ripe, and its up to you if these are just temporary placeholders while you rebuild the RM DCs as "vanilla" DCs or if you want to just put new DCs into place anyway and leave them.

 

Last but by no means least, don't forget that an XP --> windows 7 migration is a complex beast in and of itself, never mind all this extra detail! The impact of having to create new software installs, validate your current software library and obtain upgrades/replacements for stuff that won't run under windows 7, the direct impact of the migration (e.g. time spent schlepping around your site rebuilding machines with the new image and testing them), user training issues, etc. is a major project in its own right.

[Roberto]

How are you going to save money moving away from an existing CC3 network, where all licenses are still valid?

 

You're right here. There's no cost saving in terms of licences in moving from CC3 to a "vanilla" XP network, if nothing else changes. You would however save the cost of a RM CAL if you combined this with moving to a much larger XP user base.

 

The real potential for saving is what happens when you upgrade your network from Windows 2003/XP to 2008r2/win7 because those CC3 licences won't help you there if you go with CC4.

 

Your investment in CC licenses will have been wasted.

 

In economics and business decision-making,

sunk costs are retrospective (past) costs that have already been incurred and cannot be recovered

. Sunk costs are sometimes contrasted with prospective costs, which are future costs that may be incurred or changed if an action is taken. Both retrospective and prospective costs may be either fixed (that is, they are not dependent on the volume of economic activity, however measured) or variable (dependent on volume).

 

In traditional microeconomic theory, only prospective (future) costs are relevant to an investment decision.

Traditional economics proposes that an economic actor not let sunk costs influence one's decisions, because doing so would not be rationally assessing a decision exclusively on its own merits.

The decision-maker may make rational decisions according to their own incentives; these incentives may dictate different decisions than would be dictated by efficiency or profitability, and this is considered an incentive problem and distinct from a sunk cost problem.

 

from:

Sunk costs - Wikipedia, the free encyclopedia

 

The costs of moving to CC3 have already been accounted for. They shouldn't influence future decisions, beyond any "upgrade" reductions in CC4 licensing due to any upgrade pricing vs. fresh install pricing being taken into account when researching the price of a CC4 upgrade.

 

Do you use RM support? If not, don't pay for it. Just think of all the third-party software (and your time!) needed to replicate everything you already have included with CC3.

 

There is a considerable cost in going vanilla from this, but then there still is a cost to going to CC4 - retraining, etc.

 

We estimate the cost to us, of going from CC3 to vanilla as part of our CC4 upgrade, as being x at the most (that assumed we purchased KACE tools for imaging, which we have now chosen not to do). Our initial estimate of the likely cost of CC4 was 2*x. RM's actual quote to do the upgrade was 3*x. As a college rather than a school we may not be a typical RM customer and we may not be typical of a lot of people here, fair enough. But the financial implications here were something we could not ignore.

Edited June 3, 2011 by Roberto

[CyBeRkId2002]

Wow... thanks again for all the advice, tips and... erm... warnings!

 

I have actually begun the process this week of recreating our GPO to what we would like when the transition is complete. After a few hiccups I believe we are nearly their and so begins the long process of testing, looking for loop-holes etc. I have also built a machine with software similar to what we currently have at present. I cant believe the difference in speed of both booting, logging on and general use and has definately confirmed to me we are moving in the right direction.

 

I have also had a play with our SAN and having successfully configured an additional server as a host, setup a lun and attached it to the server I am now a lot calmer on that front!

 

Our final really major decision is on how we are going to deploy the OS and Software. I am completely at a loss here and something I will have to do much more research in to. Initially I was drawn to RIS/WDS with individual MSI's being allocated but am now considering imaging complete network builds.

 

Finally, could someone please explain the costs people are talking about when moving from CC3 to vanilla. We only use one or two bits of the RM management system, and plan to work in house on scripts etc for account creation etc. so am unsure what costs people are referring to (other than our time).

 

Finally, to the people who have moved (or are in the process of moving from RM) am I right in thinking I can (in our 3 DC setup):

 

• Demote server 2 as a DC
  
• Wipe and Rebuild
  
• Promote to DC
  
• Demote server 3 as a DC
  
• Wipe and Rebuild
  
• Promote to DC
  
• Transfer FSMO from server 1 to 2 or 3
  
• Demote server 2 as a DC
  
• Wipe and Rebuild
  
• Promote to DC
  

 

Thanks again,

Michael

[irsprint84]

DOnt know if its me but WDS was a pain! should defo look at fog

[Roberto]

Finally, could someone please explain the costs people are talking about when moving from CC3 to vanilla. We only use one or two bits of the RM management system, and plan to work in house on scripts etc for account creation etc. so am unsure what costs people are referring to (other than our time).

 

Well the costs are replacing those things you use in the current setup. If you don't use much of CC3 (like us) then the cost to replace the functionality you use will be minimal. But one way or another, they exist.

 

And don't discount the importance of the cost of your time. That's the one thing you can't really recover later if it turns out that things aren't going as smoothly as you hoped.

[irsprint84]

Well the costs are replacing those things you use in the current setup. If you don't use much of CC3 (like us) then the cost to replace the functionality you use will be minimal. But one way or another, they exist.

 

And don't discount the importance of the cost of your time. That's the one thing you can't really recover later if it turns out that things aren't going as smoothly as you hoped.

 

If I were to start something like this I would start now, 12 weeks of mostly low time in schools up to September

[Roberto]

If I were to start something like this I would start now, 12 weeks of mostly low time in schools up to September

 

We're starting now and plan to go live in september 2012, but we have a very complex set of questions to answer in our migration and not as many resources as we would like to answer them.

[bart21]

For imaging I would seriously look at fog.

[Boredguy]

For the software deployment, I'd look at WPKG as you can use your current CC3 packages when you create the xml files, and still allows the granular deployment, as well as some more snazzy options.

 

For imaging, the MDT2010 was quite easy to setup, and if you use the database option within it you can enter the computers mac or uuid address so that subsequent rebuilds are more automated.