Room Finder Blank if ampersand ‘&’ in room mailbox name

We have kicked off a project to try and get rid of public folders, and as such we are creating room mailboxes for all our meeting rooms.

We have found an issue in Outlook 2010 where if a room mailbox has an ampersand in its name i.e. R&D. The room list will be blank and the room will not add & hence you can not book.

The problem does not occur on outlook 2003 or 2007 and appears to be caused by the ‘room finder’ feature in outlook 2010.

We have a lot of existing room names and resources that currently contain ‘&’ in their names so would like to keep these in the new room and equipment mailboxes to make things consistent.

Waiting on a fix from Microsoft.

Sync Issues in Outlook / Exchange 2010

Our environment:

All users are on Exchange 2010, using a mix of outlook 2003-2010, with cached mode.

Server infrastructure is 4 CAS/HUB servers, and 4 mailbox server’s with a DAG.

We have decommissioned all our existing exchange 2003 servers, bar one.

So we are running in mixed mode exchange 2010, exchange 2003, purely because we need to have one exchange 2003 server on as a workaround for an organizational forms issue with Exchange 2010(see details below).

Naturally this server is just wasting power, and provides unnecessary overhead, regarding patching etc., so we would like to get rid of it however without it clients get the error below.

i.e. when we make an exchange 2010 server the replica for the organizational forms.

Sync Issues in Exchange / Outlook

You receive an error message when you try to synchronize your offline folder file if the Organizational Forms Library on your Exchange computer has been re-created or cannot be located

8:15:49 Synchronizing Forms

8:15:49 Downloading from server ‘XXXXXXXXX’

8:15:49 Error synchronizing folder

8:15:49                   [80004005-501-4B9-560]

8:15:49                   The client operation failed.

8:15:49                   Microsoft Exchange Information Store

8:15:49                   For more information on this failure, click the URL below:

8:15:49                   http://www.microsoft.com/support/prodredirect/outlook2000_us.asp?err=80004005-501-4b9-560

It appears to be a known bug, with a fix supposedly due in exchange 2010 SP1, but we are past that and still no news..

Some suggestions by MVP’s etc to just tell users to just ignore the sync issue is not an option for us, as it would still cause a lot of user confusion and calls to our internal helpdesk.

Some more references to the problem from other people:

http://busbar.blogspot.com/2010/04/exchange-2010-bug-sync-issues-with.html

http://social.technet.microsoft.com/Forums/en-US/exchange2010/thread/e2ca3480-0855-4822-bc42-3747a4bd1b1d/

And this link however I don’t have access to the post http://social.microsoft.com/Forums/en-US/partnermsgexchange/thread/cf6fc702-cfed-4818-8b51-47b6042f8329

Some more detials on the issue:

• We see the problem in all versions of outlook(2k3 to 2010)
• All our 2010 exchange servers are running Sp1 Update rollup 3v3, we have held off going to roll up 4 due to the many issues experienced with rollup 3, and the bugs that were introduced in 4.
• The DAG is across a dark fibre link connecting our Prod Datacentre to our DR Datacentre.
• We have a separate storage team who have run diagnostics and say the physical disk infrastructure is fine. It is running on IBM XIV

We have had no solutions from Microsoft thus far on this issue, so anyone that has any feedback it would be greatly appreciated.

Outlook Web App didn’t initialize error after Ex2010 SP1 Update Rollup 3

Environment:

4x Win2k8 R2 CAS/HUB Servers

4x Win2k8 R2 Mailbox Servers

All servers were installed with Update Rollup3 from a powershell window run as admin. i.e. Administrator was in titlebar

On all servers UR3 was installed with the following paramaters E:\Exchange2010Scripts\Exchange2010-KB2529939-v3-x64-en_UpdateRollup3.msp /lxv* c:\SP1UR3.log

The rollup on all servers stated that it finished sucessfully in the log file, the event log and through the installer.

However 1 out of the 4 CAS servers had a serious OWA issue:

 

Outlook Web App didn’t initialize. If the problem continues, please contact your helpdesk.

Couldn’t find a base theme (folder name=base).

Checking the C:\ExchangeSetupLogs\Updatecas.log showed the reason for this error..

  [15:45:24] ***********************************************
[15:45:24] * UpdateCas.ps1: 15/10/2010 3:45:24 PM
[15:45:24] Error updating OWA/ECP: The term 'Get-ExchangeServer' is not recognized as the name of a cmdlet, function, script file, or operable program. Check the spelling of the name, or if a path was included, verify that the path is correct and try again.
[16:59:55] ***********************************************
[16:59:55] * UpdateCas.ps1: 15/10/2010 4:59:55 PM
[16:59:55] Error updating OWA/ECP: The term 'Get-ExchangeServer' is not recognized as the name of a cmdlet, function, script file, or operable program. Check the spelling of the name, or if a path was included, verify that the path is correct and try again.
[20:33:15] ***********************************************
[20:33:15] * UpdateCas.ps1: 4/05/2011 8:33:15 PM
[20:33:15] Error updating OWA/ECP: The term 'Get-ExchangeServer' is not recognized as the name of a cmdlet, function, script file, or operable program. Check the spelling of the name, or if a path was included, verify that the path is correct and try again.
[22:59:19] ***********************************************

As can be seen this same issue occured on previous update rollups on this server.

To fix the problem i manually re-ran the C:\Program Files\Microsoft\Exchange Server\V14\Bin\UpdateCas.ps1 script, and now the updatecas.log shows it sucusfully updated owa/ecp directories, and everything was working again.

My question is why does this server keep doing this, and i have seen numerous people post questions about the same issue.

I looked at the updatecas.ps1 script and there is a line to add the Exchange snap in:

add-PSSnapin -Name Microsoft.Exchange.Management.PowerShell.E2010

So why in some instances is this not working and why isnt this error being captured by the update rollup?

UPDATE: Thanks to IAMEdward from technet forums we have a solution & a explanation for why it is occurring!

The reason this is happening is there are some Powershell registry keys missing in [HKEY_LOCAL_MACHINE\SOFTWARE\MicrosoftPowerShell\1\PowerShellSnapIns].  In my case Microsoft.Exchange.Management.PowerShell.E2010.reg and Microsoft.Exchange.Management.Powershell.Support.reg were missing on the Hub/CAS server that was throwing this error in OWA.  I exported the registry keys from one of my other CAS servers and then imported them into the one that was missing the keys.

I can confirm the CAS server i was having the issue with, also had the Powersehell.E2010 and Support reg keys missing under [HKEY_LOCAL_MACHINE\SOFTWARE\MicrosoftPowerShell\1\PowerShellSnapIns].

When i ran get-pssnapin -registered i only had the Microsoft.Exchange.Management.PowerShell.Setup.

So I was missing

Microsoft.Exchange.Management.PowerShell.E2010
Microsoft.Exchange.Management.Powershell.Support

Exporting the REG from another CAS and loading it in the broken CAS, looks to have solved the probelem as i can now run Add-PSSnapin Microsoft.Exchange.Management.PowerShell.E2010 without error.

This was not picked up because any exchange commands were always run from the Exchange management Shell, which loads  ‘E:\Exchange Server\V14\bin\RemoteExchange.ps1′; Connect-ExchangeServer -auto” which obviously uses a different method to connect(than the standard add-pssnapin), which worked even when the above registry keys were not there.

Also it would seem that the error checking in the Exchange installers is not very good, as all install log files exit sucessfully with a zero exit code.

So might be a good idea for MS to put some better error checking in, and at the very least verify the above reg keys exist.

Cheers All

SCCM OSD and MDT 2010 issues with powershell role.

The task says ‘Note:  Items supported by each operating
system are different; those that are not supported will be ignored."
However some features if selected such as "Windows Powershell" will
cause the entire roles and features step to fail, this appears to be a bug with
SCCM/MDT.

 

Battery Cache Issues with HP ML350 and SCCM OSD

Our image is deployed using SCCM’s OSD, it is a wim image of Windows 2003 SP2 R2, we
inject the drivers at build time into the build.

Without the battery, we can rebuild over and over with no issue, even if we blow away the
array.

However if we plug the battery in, and delete the array and then recreate it as a
raid 1+0, then attempt to deploy the build again we get the error below.

We have tested on multiple servers and using different batteries..

USB

If you boot WINPE from a USB device, there is a built-in SCCM step that
copies WinPE to the local hard drive and then restarts the server using the
WinPE that was loaded on the HD.

This is deploying winpe2.0/vista pe boot loaders and mbr to the HD
(NOTE: Windows 2003 has a different boot loader, mbr to vista/win7/win2008).

Then it deploys the OS to the HD, and must reconfigure the boot loader,
mbr etc for the specific OS you are deploying. (if this is anything older than
Win2008 it must change the boot loader mbr etc).
It looks to me like the process of loading Winpe to HD then deploying OS is
causing the issue. Possibly something to do with the batter cache not flushing
correctly or controller not returning correct information to the PE
environment.

DVD

For a DVD build winpe will boot from the DVD then the OS will be
deployed straight to c:\, the step of copying winpe to the local hard drive
does not occur.

I preformed a lot of tests myself before escalating this issue to HP, in
an attempt to narrow down the issue, the results are below.

 Test Results G6 Cache issue USB build SCCM, HP Server has latest firmware

Battery Cache	Rebuild Array	OS Deployed	Result	Comments
Test1         yes	yes	w2k3	Fail
Test2         no	yes	w2k3	Success
Test3         yes	no	w2k3	Success	Had existing build on it from test2, then server was powered off and cache plugged in.
Test4         yes	yes	w2k3	Fail
Test5         yes	no	w2k3	Success2	try rebuild after previous failure, don’t rebuild array
Test6         yes	yes	w2k3	Fail
Test7         yes	yes	win2008	Success	Deploying windows 2008 works fine, with same build using sccm usb build.

UPDATE1:

 The HP Level 3 believe they have identified a potential cause for the problem. A beta firmware is being written and tested to see if it resolves the problem. I expect further news next week some time.

 Update2:

HP have fixed the issue in their new version of firmware

SCCM OSD BOOT ISSUE HP ILO

SCCM OSD Intermittent issue with HP ILO, mounting
remote
ISO.

 

This error seems to occur quite frequently when using SCCM with HP ML350 with ILO 2.

From the logs below, it would appear that winpe checks for the existance of the CDROM containing the SCCM build media before all virtual drivers have been completley loaded.

The diskpart clearly shows that there is two removable devices D and E, E being the virtual iso mounted through the ILO interface and the drive with the SCCM disk in it. However the logs from sccm below show it fails because it only looks in D:\ for SCCM media, presumably because the virtual cdrom hasn’t been loaded yet. If i press F8 to load CMD prompt then launch x:\sms\bin\x64\tsbootshell.exe it goes through fine every time and when viewing the smsts.log in this scenario shows searching for SCCM media in D and E drives (see logs below).

Microsoft
DiskPart version 6.1.7600

Copyright
(C) 1999-2008 Microsoft Corporation.

On
computer: MINWINPC

 DISKPART>
list volume

   Volume
###  Ltr  Label        Fs     Type        Size     Status     Info

 
———-  —  ———–  —–  ———-  ——-  ——— 
——–

  Volume
0     D                       DVD-ROM         0 B  No Media

  Volume
1     E   SCCM         CDFS   CD-ROM       196 MB  Healthy

 

SMSTS.log

<![LOG[Checking
for
"D:\SOURCES\BOOT.WIM"]LOG]!><time="10:48:10.336+000"
date="01-19-2010"
component="TSBootShell"
context="" type="0" thread="828"
file="configpath.cpp:107">

<![LOG[Failed
to
find the configuration path.

The
system cannot find the file specified. (Error: 80070002; Source:
Windows)]LOG]!><time="10:48:10.352+000"
date="01-19-2010" component="TSBootShell"
context="" type="3" thread="828"
file="bootshell.cpp:550">

<![LOG[Execution
failed
with error 80070002.]LOG]!><time="10:48:10.352+000"
date="01-19-2010" component="TSBootShell" context=""
type="3" thread="828"
file="bootshell.cpp:634">

 

Pressed
F8 to launch CMD prompt then ran x:\sms\bin\x64\tsbootshell.exe

 SMSTS.log

<![LOG[Checking
for
"D:\SOURCES\BOOT.WIM"]LOG]!><time="10:50:34.980+000"
date="01-19-2010"
component="TSBootShell"
context="" type="0" thread="1084"
file="configpath.cpp:107">

<![LOG[Checking
for
"E:\SOURCES\BOOT.WIM"]LOG]!><time="10:50:34.980+000"
date="01-19-2010"
component="TSBootShell"
context="" type="0" thread="1084"
file="configpath.cpp:107">

<![LOG[Found
E:\SOURCES\BOOT.WIM]LOG]!><time="10:50:36.431+000"
date="01-19-2010"
component="TSBootShell"
context="" type="0" thread="1084"
file="configpath.cpp:116">

<![LOG[Found
config
path E:\]LOG]!><time="10:50:36.431+000"
date="01-19-2010" component="TSBootShell"
context="" type="1" thread="1084"
file="bootshell.cpp:553">

<![LOG[Booting
from
removable media, not restoring bootloaders on hard
drive]LOG]!><time="10:50:36.431+000"
date="01-19-2010" component="TSBootShell"
context="" type="1" thread="1084"
file="bootshell.cpp:564">

SCCM OSD ISSUES

I am trying to build a HP x64 2008 R2 server using SCCM SP2 OSD with
MDT2010.
I have customized the task sequence, which deploys the OS
fine.
The issue i am having is after the Setup windows and ConfigMgr
step, i install roles and features, then want to do a reboot to apply
the roles etc because subsequent steps require some of the roles to be
installed before they will run, i.e. HP PSP(psp-8.30.w2k8R2.x64.exe
/install) requires SNMP role to be installed. <Info on HP PSP can be found Here. >

But after the reboot
step the task sequence fails and doesn’t resume.
Is there any
pre-requisites needed if we want to put a reboot in an OSD step?

i.e. Gather, Use toolkit steps, before or after reboot? The reboot is
set to  “Currently installed default OS”.

I have tried numerous
things..
If i put a reboot anywhere in my task sequence(after the Setup Windows and Config Mgr step), the steps following the reboot do not run,
no advertisements apply post build and i have the issue specified here
http://social.technet.microsoft.com/Forums/en/configmgrosd/thread/694df987-3278-424c-bdb2-0aa83b315e1f?prof=required
i.e. SCCM is stuck in provisioning mode.

If i take the reboots out the TS completes just fine and everything is
ok(e.g. not stuck in provisioning mode, and all steps get fired)..
However i need reboots for 2 reasons:
1. because subsequent steps needs the roles installed e.g HP PSP and
snmp role.
and
2. To avoid the broadcom teaming bug with Hyper-V.. http://www.mvug.co.uk/blogs/mvugblog/archive/2009/01/25/broadcom-nic-teaming-and-hyper-v.aspx
e.g. i need to have the hyper-v role installed and thus a reboot before
i do the teaming step.

It seems to me like SCCM is having an issue resuming from the reboot and carrying on the remainder of the TS with the SCCM service installed.

i.e. without the reboot SCCM is running in a different context in that it is running without sccm service being installed, but after the reboot it switches over, as can be seen when viewing the logs, i.e. in winpe mode without SCCM service installed it is logging to smsts.log under x:\windows\temp\smsts\smsts.log, then after reboot it collates all logs and starts logging to c:\windows\syswow64\ccm\logs\ as it would under normal operation in full OS.

See below for
OSD steps.

The Task Sequence..

Full Size

 

 Full Size

After
reboot Loads to WINPE and says initializing SMS client… Then just reboots(log files below show that it gets an error and never initializes).

 

Log Files

Smsts.log

Failed to set log directory. Some
execution history may be
lost.

The system
 cannot find the file specified. (Error: 80070002; Source:
 Windows)        OSDSetupHook        19/01/2010
 9:24:41 AM        912 (0x0390)

Executing
 task
 sequence        OSDSetupHook        19/01/2010
 9:24:41 AM        912 (0x0390)

!sVolumeID.empty(),
 HRESULT=80004005
 (e:\nts_sms_fre\sms\framework\tscore\resolvesource.cpp,467)        OSDSetupHook        19/01/2010
 9:24:41
 AM        912 (0x0390)

!sVolumeID.empty(),
 HRESULT=80004005
 (e:\nts_sms_fre\sms\framework\tscore\resolvesource.cpp,467)        OSDSetupHook        19/01/2010
 9:24:41
 AM        912 (0x0390)

!sVolumeID.empty(),
 HRESULT=80004005
(e:\nts_sms_fre\sms\framework\tscore\resolvesource.cpp,467)        OSDSetupHook        19/01/2010
 9:24:41
 AM        912 (0x0390)

!sTSMDataPath.empty(),
 HRESULT=80070002
 (e:\nts_sms_fre\sms\framework\tscore\resolvesource.cpp,1395)        OSDSetupHook        19/01/2010
 9:24:41
 AM        912 (0x0390)

Task
 Sequence environment not found.  Task sequence completed in Windows
 PE.        OSDSetupHook        19/01/2010
 9:24:41 AM        912 (0x0390)

Uninstalling
 Setup
 Hook        OSDSetupHook        19/01/2010
 9:24:41 AM        912 (0x0390)

Removing
 setup hook from registry.        OSDSetupHook        19/01/2010
 9:24:41 AM        912 (0x0390)

Successfully
 removed
 C:\WINDOWS\system32\OSDGINA.DLL        OSDSetupHook        19/01/2010
 9:24:41 AM        912 (0x0390)

Could not
 delete the file C:\WINDOWS\system32\OSDSETUPHOOK.EXE. Error code
 5        OSDSetupHook        19/01/2010
 9:24:41 AM        912 (0x0390)

Marking the
 file C:\WINDOWS\system32\OSDSETUPHOOK.EXE for deletion on
 Reboot        OSDSetupHook        19/01/2010
 9:24:41 AM        912 (0x0390)

Successfully
 removed
 C:\WINDOWS\system32\OSDSETUPHOOK.EXE        OSDSetupHook        19/01/2010
 9:24:41
 AM        912 (0x0390)

Successfully
 removed
 C:\WINDOWS\system32\_SMSOSDSetup        OSDSetupHook        19/01/2010
 9:24:41 AM        912 (0x0390)

Finalizing
 logging from process
 908        OSDSetupHook        19/01/2010
 9:24:41 AM        912 (0x0390)

Successfully
 finalized logs to SMS client log directory from
 C:\WINDOWS\SysWOW64\CCM\Logs        OSDSetupHook        19/01/2010
 9:24:41 AM        912 (0x0390)

Cleaning up
 task sequencing logging
 configuration.        OSDSetupHook        19/01/2010
 9:24:41 AM        912 (0x0390)

 Execmgr.log

Common
Client Agent Settings for the client are missing from
WMI.        execmgr        19/01/2010
9:26:19 AM        3496 (0x0DA8)

As you can see from my Task sequence steps above i have set it to log progress to a progress.log, showing time and step its upto.

e.g. cmd /c echo %date% %time% Before reboot >> c:\admin\install\logs\Progress.log

As can be seen by viewing the progress log below sccm never executes the next step in the OSD after the reboot step.

Progress.log

Tue
19/01/2010  9:17:19.74 Before reboot

Stuck in Provisioning Mode

PS C:\Users\Administrator> Reg query HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Microsoft\CCM\CcmExec

HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Microsoft\CCM\CcmExec
CheckIdleEndpointSeconds    REG_DWORD    0x12c
CheckMemorySeconds    REG_DWORD    0x3c
FileCleanupSeconds    REG_DWORD    0xa8c0
CoFreeSeconds    REG_DWORD    0x258
LowMemoryThresholdMBytes    REG_DWORD    0x14
SystemTaskExcludes    REG_SZ    SchedulerStartup;SchedulerShutdown;SchedulerLogon;SchedulerLogoff;ClientRegistrationStartup
CheckUserSeconds    REG_DWORD    0x258
ProvisioningMode    REG_SZ    true
ADTimeOutSeconds    REG_DWORD    0x258

HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Microsoft\CCM\CcmExec\Commands

UPDATE!!

Looks like its one of the roles and
features i have selected… if i
disable the roles and features step, my TS completes fine, if its
enabled and a reboot is included afterward it never completes.

Here is an export of the roles and features i have set…

 <step type=”BDD_InstallRoles” name=”Install Roles and
Features”
description=”” runIn=”WinPEandFullOS” successCodeList=”0″>
<action>smsswd.exe /run: cscript.exe
“%DeployRoot%\Scripts\ZTIOSRole.wsf”</action>
<defaultVarList>
<variable name=”OSFeatures”
property=”OSFeatures”>NET-Framework,NET-Framework-Core,NET-Win-CFAC,NET-HTTP-Activation,NET-Non-HTTP-Activ,NET-Framework,NET-Framework-Core,NET-Win-CFAC,NET-HTTP-Activation,NET-Non-HTTP-Activ,GPMC,RSAT-ADCS-Mgmt,RSAT-ADDS,RSAT-SNIS,RSAT-ADLDS,RSAT-AD-Tools,RSAT-ADDS,RSAT-ADDS-Tools,RSAT-AD-AdminCenter,RSAT-SNIS,RSAT-ADLDS,RSAT-AD-PowerShell,RSAT-DHCP,RSAT-DNS-Server,RSAT-File-Services,RSAT-DFS-Mgmt-Con,RSAT-FSRM-Mgmt,RSAT-NFS-Admin,RSAT-Hyper-V,RSAT-Print-Services,RSAT-RDS-RemoteApp,RSAT-Web-Server,RSAT-Hyper-V,SNMP-Services,SNMP-Service,SNMP-WMI-Provider</variable>
<variable name=”OSRoleServices”
property=”OSRoleServices”>AS-Web-Support,FS-FileServer,FS-DFS,FS-Resource-Manager,Print-Server,Web-WebServer,Web-Http-Errors,Web-Asp-Net,Web-Net-Ext,Web-Http-Logging,Web-Log-Libraries,Web-Request-Monitor,Web-Http-Tracing,Web-Basic-Auth,Web-Windows-Auth,Web-Mgmt-Tools,Web-Mgmt-Console,Web-Scripting-Tools,Web-Mgmt-Compat,Web-Metabase,Web-WMI,Web-Lgcy-Scripting,Web-Lgcy-Mgmt-Console</variable>
<variable name=”OSRoles”
property=”OSRoles”>Hyper-V</variable>
</defaultVarList>
</step>

Checking the ztiroles log shows…

1344:
2010-01-07 15:15:42.039 [Sync] Warning (Id=0) Sync of ‘.NET Framework
3.5.1’
skipped because the sync of its antecedent ‘.NET Framework 3.5.1’ needs
reboot

1344:
2010-01-07 15:15:42.039 [ServerManagerCmd] Error (Id=0) Installation of
[Application Server] .NET Framework 3.5.1 failed. The computer must be
restarted to complete another installation. After restarting the
computer,
please re-run the same command to install this role.

1344:
2010-01-07 15:15:42.086 [Sync] Warning (Id=0) Sync of ‘HTTP Activation’
skipped
because the sync of its antecedent ‘HTTP Activation’ needs reboot

1344:
2010-01-07 15:15:42.086 [ServerManagerCmd] Error (Id=0) Installation of
[Application Server] HTTP Activation failed. The computer must be
restarted to
complete another installation. After restarting the computer, please
re-run the
same command to install this role.

1344:
2010-01-07 15:15:42.086 [Sync] Warning (Id=0) Sync of ‘Web Server (IIS)
Support’ skipped because the sync of its antecedent ‘Static Content’
needs
reboot

1344:
2010-01-07 15:15:42.086 [ServerManagerCmd] Error (Id=0) Installation of
[Application Server] Web Server (IIS) Support failed. The computer must
be
restarted to complete another installation. After restarting the
computer,
please re-run the same command to install this role.

1344:
2010-01-07 15:15:42.086 [ServerManagerCmd]          [STAT]

Overall Sync Time:
‘33.8820381’ second(s)

1344:
2010-01-07 15:15:42.086 [ServerManagerCmd]          Error:

Installation of one or more
roles, role services, or features failed, and a restart is required.
Check the
Server Manager log for additional information and try again.

Looking
at the log file above, i un-ticked the Web Server (IIS) and HTTP
Activations(see below).

SCCM
OSD package now reboots and resumes successfully, and the features that
i unselected are in fact still installed, so another feature must be
installing the features as part of their prerequisites, issue
finally solved!!

So in the end it was due to the Roles and Features Step, it would seem that when a reboot was placed in the TS, and there was a feature(s) that had failed to install because it needed a reboot, when this reboot happened(initiated by the ts)  it would try and install the features, on boot and then reboot underneath SCCM hence leaving SCCM in an unknown or provisioning state.

That probably didn’t make any sense so ill try to explain by means of an example.

In Roles and Features, the Web Server (IIS) Support feature when selected failed to install with a “The computer must
be
restarted to complete another installation. After restarting the
computer,
please re-run the same command to install this role.” 

After the Roles and Features step, i have a reboot in my TS.

The TS initiates this reboot fine, when the server comes up from this reboot, SCCM shows initializing the config mgr client….

In the back ground it would seem that the Web Server IIS Support feature is now trying to install.

After a few mins the config mgr client initializes and sccm resumes the ts..

Meanwhile the Web Server IIS Support feature is still installing.

So the SCCM ts could now be part way through its remaining steps and all of a sudden the server reboots again (unexpectedly), i.e. not initiated by sccm or the ts, so my only guess is that the Web Server IIS feature has completed and is forcing the reboot itself.

Then when the server boots up sccm is stuck in provisioning mode, and never completes all its remaining steps.

Any feedback welcome, i can’t see the internal workings of SCCM so the above is just my observations\thoughts.