Rocket Man

Hey just seen this post.... I have just created a fully blown SCCM 2012 SP1RTM production infrastructure over the past 3 days. This was not an upgrade from non-SP1 it was a brand new install. I am having no issues with OSD and client installations. It works both via OSD and client push method. The clients are reporting back as normal as I can do all the normal stuff as usual like hardware inventories, RMC etc. Is this an issue that only occurs from upgrading an NON-Sp1 infrastructure to SP1 can anyone verify? I would be keen to know as I have 5 other non-Sp1 sites that I am considering upgrading. Thanks

You could add the WinPE3.0 to your boot images...this guarantees successful boots of an extensive range of Dell models!

Just use the associate MAC address to system report!! But the successful PCs that get imaged will no longer be seen as unknown they will get the name stamp you have specified during OSD. Thus the only unknown system will be that of your failed PC!! Just delete this account(unknown) and away you go again!! Rocket Man

Have you tried plain old collection variables? (Not sure if it will work but it's worth a try!!) If you have your SCCM collections divided into 2 collections i;e laptop collection and Desktop collection, you could place a collection variable in both collections i:e LAPTOP Collection: Install Laptop Software = true DESKTOP Collection: Install Desktop Software=true Then on your TS you could have 2 install applications tasks grouped to segregate the models i:e The 1st application task in your TS could have a group called laptop software with the appropriate applications specified for laptops and then on the actual group folder specify a condition Install Laptop Software = true The 2nd application task in your TS could have a group called desktop software with the appropriate applications specified for desktops and then on the actual group folder specify the condition Install desktop software = true The theory behind this is when you deploy your TS to both collections when it comes to the time of installing the applications it will check against the variables set, so if it is the laptop collection it will install the laptop grouped applications and vice versa!! Give it a go and see Again I am not sure if this works!!

The registry should execute if you have all the registry enteries in the one registry key called regedit.reg, if all keys are specified in this and where to add to the registery it should work!! Works well for me on a specific piece of software!! As for your other problem .....well atleast the files are now copying to the local machine, unfortunately I can give you no more assistance on this area as I only run x86 clients so have no idea as to why they are copying to the sysWOW64 folder?? Maybe some one else knows and would like to contribute? Rocket Man

What is the syntax of your bat file at present?

Hi peter this is only for new machines that come into the SCCM database!! The collection the custom software updates poilcy is targeted at is only for new machines, after the 1st day of been known to SCCM it drops back out of the collection and then recieves the normal default settings or in your case your adjusted 12 hours cycle!! Use at your own will.......there will of course be extra churn on the network if you are using unknown deployment for the 1st time, but after this use staged deployment and have your default policy target the proper collections that the machines are in and they will pull down the policies at the same speed as the scenario above as they are not new machines they just get a new ID and immediately know what policies the have associated with them!! I have tested this in many many labs and it works seemlessly, once the unknown machine becomes known it joins the 1 day collection as I call it and recieves all software updates immediately including the FEP updates, then after the day it drops back out of the collection thus leaving no churn on the network all automated with the query on the collection of course!! Rocket Man

Have a package with your files in it and a batch file also: The syntax so it deploys from the DP ---------------------------------------------------------------------------------- @Ehco copy %~dp0filename1.extension "c:\windows\system32" copy %~dp0filename2.extension "c:\windows\system32" copy %~dp0filename3.extension "c:\windows\system32" copy %~dp0filename4.extension "c:\windows\system32" copy %~dp0filename5.extension "c:\windows\system32" regedit.exe /s regkeys.reg @Exit --------------------------------------------------------------------------- your rekeys.reg would have to be taken from a machine with the app installed!! Put all the keys into one AKA regkeys.reg Rocket Man

Maybe clearing the last pxe deployments on the collections would of worked too!

what are your APP catalog components states reporting? Are they Healthy?

Just to confirm I have implemented this service on a test lab. It was an easy installation of the webservice as per the instructions. I have to say it does work extremely good, as I have tested it out quite a bit on the test lab which consists of 4 test VMs to deploy to. It auto-named them seemlessly and placed them in the target OU specified!! I would recommend anyone who is looking for an auto_naming service to definitely try this out. Thanks rpieniaz Rocket Man

Are you deploying a new OS via SCCM to these new machines initially or are you just joining them to the domain? To get the client installed straight away, use unknown computer support, build your custom Task sequence with the configMgr client attached. deploy it to the unknown computer collection(which will be any new machines that is not known to SCCM) and make it available not mandatory. PXE boot up your machines and choose the TS. This way when ever your machines are built, they will be autonmatically joined to the domain via the configuration in the TS and they will also have the client installed. AS for the endpoint installation take a look at this: http://blogs.technet.com/b/michaelgriswold/archive/2012/08/02/fast-policy-evaluation-only-for-new-machines.aspx This works brilliant for new machines and they recieve policies almost instantly depending on the time you configure in the custom software evaluation and software deployment schedule.........i use 5 mins and 10 mins for them!!

I will give this a go!!

Not sure what you mean here? Are you not targeting machines with the client install either via Client push or update? Are you waiting for the automatic site wide method? Have you targeted collections with the endpoint policy and anti-malware policy? The default client polling time is 60 mins for policies, I would suugest you change this if you'd like to have client machines pole for new policies quicker!!

Hi Niall it was a manual install for both instances then I installed the CU2 on top of this which successfully updated both instances. Now the second instance for WSUS is it also a prereq for the collation to be SQL_LATIN1_GENERAL_CP1_CI_AS? My thinking is YES as this is the collation for a single instance hosting both DBs. For all SQL roles i specify a single domain account just to keep it simple. Just to add I did set this implementation up using the system account for all roles and even this had the same outcome but if I use just one instance for both DBs it has no errors and works with both the domain account credentials for all SQL Roles\service accounts and also works with the system account. I also will add I have 4 seperate smaller SCCM 2012 non SP1 sites sitting on server 2K8R2 SP1 and SQL Server 2K8 Sp1 CU6 with 2 instances using domain credentials and these have no problems. EDIT: it is CU3 that i am using and not CU2 as mentioned Thanks

OK......just thought I would bring this thread back up. With the release of SP1 RTM yesterday I set up a testing lab last night with the same set up as before SCCM 2012 SP1 RTM with Server 2012 and SQL2012 CU2 with 2 instances, one for the ConfigMgr DB and one for WSUS. Could not believe it, the same error again as above with the 2 seperate instances. Very frustrating to say the least as both instances are setup the same way as I would set up a SCCM implementation on a single instance, using a domain account for all SQL roles etc......and after reverting my Beta lab to a single instance after the 1st error and seeing this worked great with no errors even made it more frustrating. Anyway as Peter suggested it was a security issue. After digging arround in the SQLWSUS log folder I came across an ERROR log and in it was the reason: 2013-01-04 01:04:16.04 spid55 Failed to verify Authenticode signature on DLL 'e:\Program Files\Microsoft SQL Server\MSSQL11.WSUS\MSSQL\Binn\msxmlsql.dll'. 2013-01-04 01:04:30.95 Logon Error: 18456, Severity: 14, State: 38. 2013-01-04 01:04:30.95 Logon Login failed for user 'NT AUTHORITY\SYSTEM'. Reason: Failed to open the explicitly specified database 'SUSDB'. [CLIENT: <local machine>] 2013-01-04 02:01:31.28 Logon Error: 18456, Severity: 14, State: 38. I never assigned the system account to the SQL setup so I do not know why it was complaining about this(especially when everything works great on one instance). What I have done(not sure if it is correct but it worked), I opened up the WSUS instance via SQL server Management studio and went to security....Logins and properties of the system account, and sure enough it was not mapped to any database, so I ticked the SUSDB on the mapping node, restarted the SCCM site server and all nice green ticks now on the site components(software update point) with no errors in the WSUS components either. Hopefully it may help some-one else in similar circumstances. Now I can go about setting up my new SCCM 2012 SP1RTM production site feeling comfortable! Rocket Man

The full download is called System center 2012 Datacenter with Service Pack 1

SCCM 2012 SP1 RTM was released yesterday for GA. You should install this to future proof yourself as it supports windows 8 and server 2012. Sql 2012 with a minimum of CU2 is supported with SCCM 2012 SP1RTM and it can be installed on server 2012. Here are the guides for SCCM 2012 on server 2k8R2 and SQL server 2008.

Yes you will. You will have to add an other Boundary group with your new siteB DP in it and add your Site B boundary IP subnet to it and remove this from the current Boundary set up!! This will service only systems at siteB thus the reason as why only siteB subnet Boundary + DP will be associated with it!! The current setup that you have will then service SiteA,SiteC and SiteD considering it is a star Topology WAN infrastructure you have? Probably better to create your second DP at Site B now, will save any alterations later to the boundaries. It may have problems or may not, I never had to change boundaries yet, but I'm sure it would be OK. EDIT: Do not forget to add your site specific collections to your Distribution point groups. You will have 2 DPGs' when ever you have your second DP created. This is all from a personal view and my own experience with my own SCCM implementations I have done to date. I am no way an expert at this but may I add I have one SCCM setup with 20 DPs, 20 Boundaries and 20 DPGs and have no problems with systems getting software from the appropriate DP assigned. You will know if your Boundaries are correct when you go to change them when you create your second DP. The systems at this AD site will lose their site code if it is not correct, if you do set it up correctly then they will keep their site code(this is one way of knowing that your boundaries are setup correctly, and the systems do not need to have the client installed for the site code to to be assigned initially).

Well sites C & D collections will have to be configured to be part of site A DPG(distribution point group) as they will have no DP available to service clients on these sites so they will be serviced via DP Site A across the WAN links. I gather you have a star topology WAN and not MESH? You will have already have 4 boundaries showing up in your Boundary node. Create 1 boundary group to start with as you only have 1 DP at present and add your DP(site server DP to this)in the references tab and add your 4 Bounadries to it in the general tab. Be much easier now for you to create your second DP at site B instead of having to alter the boundary configuration later on!! Rocket Man

Thanks Niall!

Hi Hoping some-one can advise on this. Tommorrow I am starting to implement quite a large SCCM infrastructure with SCCM 2012 SP1 RTM, Server 2012 and of course SQL Server 2012. Just wondering what is the latest SQL server version that is supported with this set up. I am currently running CU2 in my test lab of SCCM 2012 SP1 Beta. Does it support SQL Server 2012 SP1? According to the this http://technet.microsoft.com/en-us/library/gg682077.aspx it is not but this may be out-dated by now? Thanks

Have you enabled unknown computer support in your enviornment? If not I would definitley recommend you do and use this method for your new PCs, will save alot of hardship of having to import computer info in, this way all you need do is f12 the machines into PXE and choose your available Task sequences. These will be initially deployed to the built in unknown computer collection and made available not mandatory! This link gives guides on how to prompt for system name during OSD http://www.windows-n...n-manager-2012/ You can use the variable that Niall demonstrates or further on down the thread you will see a script that I posted(may I add I found this on the web, so thankyou to the owner of it). If you use the script that I posted(PromptForSystemName.vbs) then you will have to create a package out of it. Then add it to your TS as a run command line Task as seen in the screen shot below just after partion disk but before apply OS task!! This is not automated unfortunately, but as i said it only has to be done once after this any new refreshes, SCCM will hold onto the systems old name. In my working environment I could safely say that we refresh approx 6000 machines per year and maybe add anywhere between 150-250 new machines also and decomission older machines of the same amount, this works for me with little help on the ground also!

All this auto-naming stuff is great in theory but why would you want SCCM to autoname machines? Say a single department of 150 new PCs and they were all to be imaged at the same time, how would this auto-naming keep the machines in a nice sequence for we all know some machines may fail OSD for whatever reasons and this means that this machine(s) has missed it's naming stamp and some other machine gets it maybe at that far end of the office (not good)..... Dunno but I like to have some sort of order in the naming sequence so I know exactly where a specific named machine is physically located, this is achieved the old fashioned way via variable/script at OSD time to the new machine and manually entered. This only has to be done once after this just use staged deployment and it will hold onto the old name. Could you just not deploy your TS(staged deployment) to a collection with this machine/machines in it.... instead of importing them in.........this way they will hold onto their old names, of course considering the ConfigMgr client is on the machine, if they are new machines just use unknown computer support for them as above with a variable configured on the unknown collection or via a script!!

This is a merley an account that you created in AD, you have to give it a role to play, maybe you have, but if you haven't you could use it as your network access account, or you could create a new account called netaccess or something similar and use this!! Administration node, site configurations, site, configure site components along the top ribbon and select software distribution, this is where you set up your Network access account! Make sure it can connect to your share!