Try the guts of the script on an affected profile. It won’t hurt anything but might fix it.
rmdir /s /q "C:\Users\<name>\AppData\Local\Microsoft\Windows\WebCache" del "C:\Users\<name>\AppData\Local\Microsoft\Windows\WebCacheLock.dat"
Sounds familiar. Are you using copyprofile=true in your sysprep?
I use the following in a vb script that runs from an auto-logged-on Administrator as it comes out of sysprep.
Sub sProfileFixes() ' Clean up problems caused by using COPYPROFILE=TRUE Dim retry : retry = 0 Do Until retry = 10 If objFSO.FolderExists( "C:\Users\Default\AppData\Local\Microsoft\Windows\WebCache" ) Then On Error Resume Next Call objShell.Run( "cmd /c rmdir /s /q C:\Users\Default\AppData\Local\Microsoft\Windows\WebCache", 0, True ) Err.Clear retry = retry + 1 Else Exit Do End If Wscript.Sleep( 5000 ) Loop If objFSO.FileExists( "C:\Users\Default\AppData\Local\Microsoft\Windows\WebCacheLock.dat" ) Then objFSO.DeleteFile "C:\Users\Default\AppData\Local\Microsoft\Windows\WebCacheLock.dat" End if End Sub
@Tom-Elliott Okay, I’ll look into that.
Ultimately I’m looking for a best-effort recommended, second to last resort (reinstalling OS would be the last resort) solution that would tear down and pave over all that is FOG.
Sometimes tinkering, tweaking and forensic fixes just don’t do the job and you need other more drastic options in your troubleshooting arsenal; a FOG uninstaller.
But I don’t want to re-install the OS by remote. Not yet at least.
Re-installing FOG does not strip it from the server. It re-installs FOG.
It’s easy enough to yum remove virtually everything else. What about a solution for FOG and its tendrils?
You misunderstand. I’m exploring a “nuke it from orbit” solution. I want to tear the server down to its bones then build it back without physical intervention.
For one particular server I used the following.
# Remove FOG services rm /etc/init.d/FOGImageReplicator rm /etc/init.d/FOGMulticastManager rm /etc/init.d/FOGScheduler # Delete FOG database mysql -u root -p <password> drop database fog; exit # Remove FOG files rm -rf /var/www/fog rm -rf /var/www/html/fog rm -rf /opt/fog rm -rf /opt/fog.old rm -rf /tftpboot rm -rf /opt/trunkgit rm -rf /opt/trunksvn # Delete FOG system user userdel fog userdel fogproject # Delete supportive systems yum -y remove php* yum -y remove httpd # Clean & Update the OS yum clean all rm -rf /var/tmp/yum-administrator-PudGB rm -rf /var/cache/yum yum list iostat yum makecache yum grouplist yum grouplist hidden yum update -y # Reboot to Commit reboot
This is a curious request I had to stop and think about. ReFS is not bootable, so it would never contain an OS, which defeats the primary purpose of using FOG.
I’m not a fan of multi-partition imaging, so I can’t say just how important this would be to FOG users in general.
Primarily I want to as it says, totally remove everything and anything that FOG has touched or used from the latest release 1.5.7 all the way back to at least 1.2.
Not that I’m giving up on FOG right now, but I have some legacy servers cruising for a bruising, and I want to be able to rebuild as cleanly as possible, by remote.
Then of course, comes the rebuild. Installing FOG is a cinch, but of course I need to re-install those OS elements affected, and I’d prefer to not have anything at all left from them that FOG may have touched … to totally remove any doubt that something, somewhere, may have been borked by FOG or the many versions installers over the years of revisions.
I’m interested in a CentOS 7 solution.
Things can break from time to time, and given the sheer variety of … variety, it’s good to know for testing purposes if there is a cutoff for how far back one can go for testing before you reach incompatibility.
That also goes for going forward, using for example a 1.4.4 system, how far up the food chain can one go with the kernels for testing?