[nobrainer]

I apologize for keeping this thread going, I’ve been tasked to get this new image done by September 30th, and I’m at the proverbial brick wall.

I signed in with the latest and greatest image I did, and I’m thinking the postflight.sh wasn’t called / ran. I can find in the logs the firstrun.sh file that was run… but I can’t find any reference to postflight.sh, and none of the settings seem to be put into the files, I even put in Echo’s so I could (hopefully) see them running in the log files.

When I make the Pkg, I can see that postflight.sh is there, I can open Pkgmaker and under contents go to firstboot –> scripts and in the section under “Postflight” I have scripts/postflight.sh, but as I said before, it doesn’t appear to be running.

I also acquired the company’s retail disk of 10.6 and imported that to use as the main install disk.

Update: Putting the postflight.sh contents into the firstboot script seemed to run everything. The login disclaimer is there, the ARD is active, the login window is list of users, firewall is on, etc

Things I see not working –

1) System Pref –> Sharing –> Remote Login “only these users” is selected but the box is blank, and
1a) –> We can’t push apps to the machine because the management account is not working

2) System Pref –> Remote Managment All users is selected, but our local admin account (which is what we want to be the only account able to do it) is listed in the box.

I guess the last 2 items I need to resolve is those.

But that being said, I think that is proof that the postflight.sh script never ran, as a lot of those settings didn’t come through until I put in the firstboot.sh file.

[nobrainer]

setting the -int 1 worked… so now the firewall is fine… the website is fine…still having trouble with the SSH, getting it to display the ‘list of users’, and display the logon banner. The SSH is critical as I can’t run any of my casper pushes / scripts / etc w/o it. I know the Casper SSH account is there because I can sign into the box with it.

Looking at line 39, it says echo allow groups domaingroup admin…. I didn’t change anything there and our admin group is named something else, could that affect it, or is that more like an argument passing?

The Banner, i think the syntax is correct, I made a separate script file for just that that works, so I’m not really sure why that one isn’t from the postflight.

[url]http://pastebin.com/2LcVfXRY[/url] is for my firstboot.sh file, it looks like it contains the info about SSH, and ARD stuff. The firewall is commented out there due to running the post flight script.

Looking at the debug log for instaDMG, it’s no longer giving an error about the package like it was doing prior.

[url]http://pastebin.com/5rHcZNv5[/url] is the postflight file, as it stands now.

Thank you!

John

[nobrainer]

Couldn’t do “r”, but “tc” seemed to work.

[nobrainer]

Thank you so much!

[url]http://pastebin.com/EYLHd1C9[/url]

[nobrainer]

I looked at the script again, I found a missing $ in the defaults write portion of the firewall… didn’t seem to make a difference.

[code]Running Script postflight.sh…
Script Exit Code:0
Script Result: Delete: Entry, “Disabled”, Does Not Exist
2010-09-22 06:22:51.869 defaults[4997:903] Unexpected argument 2; leaving defaults unchanged.
Unmounting file server…[/code]

was all I got.

[nobrainer]

Well, I’m trying a new image… and I added some stuff to the firstboot pkg, but now I’m getting some errors…

[code] PHASE:Running package scripts…
%74.939766
./postflight: /Users/jwojda/instaDMG/InstallerFiles/InstaUp2DatePackages/Firstboot.mpkg/Contents/Packages/firstboot.pkg/Contents/Resources/postflight: line 48: unexpected EOF while looking for matching `”‘
./postflight: /Users/jwojda/instaDMG/InstallerFiles/InstaUp2DatePackages/Firstboot.mpkg/Contents/Packages/firstboot.pkg/Contents/Resources/postflight: line 51: syntax error: unexpected end of file
PackageKit: Install Failed: PKG: post-flight scripts for “com.shc.firstboot.firstboot.pkg”
Error Domain=PKInstallErrorDomain Code=112 UserInfo=0x1004b6870 “An error occurred while running scripts from the package “firstboot.pkg”.” {
NSFilePath = “./postflight”;
NSLocalizedDescription = “An error occurred while running scripts from the package U201cfirstboot.pkgU201d.”;
NSURL = “./Contents/Packages/firstboot.pkg — file://localhost/Users/jwojda/instaDMG/InstallerFiles/InstaUp2DatePackages/Firstboot.mpkg/”;
PKInstallPackageIdentifier = “com.shc.firstboot.firstboot.pkg”;
}
install:didFailWithError:Error Domain=PKInstallErrorDomain Code=112 UserInfo=0x1004b6870 “An error occurred while running scripts from the package “firstboot.pkg”.”
%97.750000
Install failed: The Installer encountered an error that caused the installation to fail. Contact the software manufacturer for assistance.
The install failed (The Installer encountered an error that caused the installation to fail. Contact the software manufacturer for assistance.)
Folder 16 Firstboot done (15:19:26)[/code]

I went through the postflight lines 48, it’s the firewall enable. I noticed the script had the globalstate -int #. However, I looked online and it looked like I could remove the -int, so I did…

[code]sudo defaults write “${PREFS_DIR}/com.apple.alf” globalstate 2[/code]

That seems to be where it’s still bombing out, I verified the Root / Wheel permissions…

[nobrainer]

[QUOTE][u]Quote by: nobrainer[/u][p][QUOTE][u]Quote by: dead2sin[/u][p]I had this issue as well. In 10.6_vanilla.catalog, comment out “Installer Disc Builds” as follows:

[code]#Installer Disc Builds = 10A432, 10B504, 10C540, 10D573, 10F569[/code]

I know there is some functionality behind that change, but I’m not quite sure what it is. I’m sure Larkost might be willing to explain it 🙂

Nate[/p][/QUOTE]

Oddly enough, I was just looking at that… I found this on Apple’s site…

[url]http://support.apple.com/kb/ht1633[/url]

there’s another build revision apparently. I re-imported the disk from a newer machine that was Disc Version 1.2 (all my previous disks are 1.0, though I tried that disk too – still didn’t work). I was just about to try plugging the newest revision numbers in to see if that made a difference, if it doesn’t then I will try your suggestion of the comment.

Thank you!

[/p][/QUOTE]

Okay, well adding the extra versions didn’t help, but commenting the whole thing out did, now the image is going!

Thank you so much!

[nobrainer]

[QUOTE][u]Quote by: dead2sin[/u][p]I had this issue as well. In 10.6_vanilla.catalog, comment out “Installer Disc Builds” as follows:

[code]#Installer Disc Builds = 10A432, 10B504, 10C540, 10D573, 10F569[/code]

I know there is some functionality behind that change, but I’m not quite sure what it is. I’m sure Larkost might be willing to explain it 🙂

Nate[/p][/QUOTE]

Oddly enough, I was just looking at that… I found this on Apple’s site…

[url]http://support.apple.com/kb/ht1633[/url]

there’s another build revision apparently. I re-imported the disk from a newer machine that was Disc Version 1.2 (all my previous disks are 1.0, though I tried that disk too – still didn’t work). I was just about to try plugging the newest revision numbers in to see if that made a difference, if it doesn’t then I will try your suggestion of the comment.

Thank you!

[nobrainer]

I should state that I am following the instaup2date guide ([url]http://www.osxdeployment.info/wiki/InstaUp2Date_Guide[/url]) to the letter, so I have multiple catalog files, and I am using the scratch disk command to launch the process. I’ve tried copy / pasting the name from the instaDMG.bash script and the FindInstallerDisc.py over the OSX install disc file name so it matches exactly. I’ve tried with and without the installerchoices.xml also…

[code]sudo ./instaUp2Date.py -p Snowleopard.catalog –instadmg-scratch-folder /Volumes/instaDMG[/code] for my external HD partition.

[nobrainer]

I’m just going to comment on #4…

Jamf Software’s Casper Suite was highly recommended to us, and we were told “a lot of companies have used other applications and then gone to Jamf’s, but they didn’t know of any that had Jamf and went to another application”

It’s a full end to end suite of applications from deployment, management, packaging, scripting, software updates, etc. It will track your any Apple inventory (iPod Touch, iPhone, iPad, etc), give you detailed data on each device, exportable inventory lists, etc.

They have excellent support, often getting back to you within minutes of asking a question to their direct support, as well as an awesome community of users (between AFP548 and Jamf’s mailing list, I have been able to get any question I can dream of answered quickly).

They integrate well with Adobe products and support Faronic’s DeepFreeze and PowerSave products natively.

They also offer services, where they will do your scripting, or packages, or images or whatever you want.

Bottom line, you get what you pay for, and I’ve yet to see a reason to move away from Casper/Jamf.

[nobrainer]

[QUOTE][u]Quote by: nobrainer[/u][p]Hmm.
the Killall command didn’t work.

$ killall -USR1 DirectoryService
No matching processes belonging to you were found

$ tail -f /Library/Logs/DirectoryService/DirectoryService.debug.log
tail: /Library/Logs/DirectoryService/DirectoryService.debug.log: No such file or directory

I can sign in as a local admin and as a domain admin. I just can’t sign in as the end user, which is what confuzzles me.

[/p][/QUOTE]

Ohh… needs a sudo killall 🙂

[nobrainer]

Hmm.
the Killall command didn’t work.

$ killall -USR1 DirectoryService
No matching processes belonging to you were found

$ tail -f /Library/Logs/DirectoryService/DirectoryService.debug.log
tail: /Library/Logs/DirectoryService/DirectoryService.debug.log: No such file or directory

I can sign in as a local admin and as a domain admin. I just can’t sign in as the end user, which is what confuzzles me.

[nobrainer]

The culprit appears to be /Library/Preferences/com.apple.Bluetooth.plist

Since I already ruled out the base image as the culprit, I did manage to find out it’s my AV of all things.

Thank you for the help!

[nobrainer]

Hmm. I’ve tried system.log, directoryservice.error.log and directoryservice.log

those 3 didn’t seem to change when I tried to sign in with the user’s account.
then I had to give up the machine back to the user.

[nobrainer]

strangely enough, I built just the base image and it wasn’t there.
However, when layered the rest of the apps on the machine for deployment to a user, the keyboard / mouse showed back up. So it DOES sound like a preference file somewhere has it in there..
however, now I have to figure out which of the other 15 or so apps is causing it.

[Author]

Posts