Yikes! Crash when declining 2nd Instance of NTLite

[pmikep]

Am running 1.6.2.6220.

My first 1.6 version. So I thought I would see how to use the Parallel Instance of NTLite.

Not finding any documentation about it, I clicked on the NTLite icon with an instance of NTLite already running and processing a base image.

I got a pop up window that explained that running in parallel was a premium feature and did I want to continue. (I have a license.)

I declined this time.

Then I got this:



It does seem as though my first instance of NTLite stopped working. It's been showing 40% for the past few minutes. My CPU isn't working hard, even though taskmanager shows ntlite using 25% and dismhost taking 5%.

Don't know what to do. When I try to cancel, I get the dire warning about leaving things in an unstable state. And I remember, from years ago, what a pain that was, trying to delete the linked files.

 

[pmikep]

This is crazy. I went back to v1.5 and my old Win7+SP1 iso. Then I integrated my standard five KB's, which were "out of order" according to v1.6.

To my surprise, even v1.5 fails when I try to load the resulting Win7+SP2 Base. This is with the NTLite directories in a different partition. (And portable installation. So in theory, nothing in my Registry to screw things up?)

The only other thing that's different since this problem beganis that I'm not letting NTLite do a Clean Up of the Updates.

So now I'm trying it again with v1.5. No Clean Up still, but I'm also not removing the other WIM's from the Base. Let's see how that goes.

 

[Clanger]

"KB's, which were "out of order". I have noticed this recently adding 3125574 to post setup, ntlite rearranged to order, i have them numbered 1 to 4, Servicing Stack, 3125574 then the other 2. It put no3 1st. I think Nuhi changed trhings to add platform updates 1st, i dont know if no3 was such an update, but my long standing set order of 1 2 3 4 has never given me a problem.

 

[pmikep]

While I'm waiting for my latest experiment to complete, here's my very simple present for what I call my latest Win7+SP2 Base. Perhaps nuhi can try to see if the resulting WIM loads for him?

 

[Clanger]

He will soon find the issue. Pray its not pebkac.

 

[pmikep]

Well, it failed again. This time it got to 41% of the load. But that might simply because I hadn't removed Editions that I didn't need.

 

[Clanger]

Take a break. Leave it for today and come back to it tomorrow with fresh eyes and you may see the problem, it will probably stare you in the face. Or Nuhi may have spotted it. A break or another pair of eyes will see what you cannot see today.

 

[Kasual]

Format the drive and start again, keep track of the log files to make reports to nuhi.

That seems like a windows/dism issue related.

 

[pmikep]

Format what drive? The drive that NTLite is installed on? The old drive that the mount folder used to be on?

 

[pmikep]

Well, this is totally hosed.

I dropped back to v1.5.0.6025 which was the last version of NTLite I used to make my current Win7 install.

I started with Win7+SP1.iso and used my old preset to bring it up to the Convenience Rollup Base. (My "SP2.") I used the Preset I had used in February.

After it was finished, I tried to mount that new Base image.

Now that mount is failing.

I don't understand. I've kind of "reverted" back to an NTLite that used to work. What's changed on my system?

Do I need to do something with/reset dism? (I don't even know where dism is.)

About the only other thing that's changed since I last did all this in February with NTLite was that I've installed a patch KB's using WSUS Offline Updater.

Is there some new patch that's preventing NTLite from mounting my Convenience Rollup image?

(I suppose, in a pinch, I could just do a Removal with an Integration of the Roll Up. But it was nice to start with the Roll Up as a Base. Didn't have to wait for the Roll up to integrate each time I tried a new Removal scheme.)

 

[Kasual]

Format what drive? The drive that NTLite is installed on? The old drive that the mount folder used to be on?

Format the OS drive.
What i really do when i got in trouble, is to install the OS without formatting.

I've read somewhere that mounting is not equal to extracting.

 

[Clanger]

If the problem is image mounting thats dism. Extract the dism files and folders from your "sp2" image and copy them to the running system.

 

[Clanger]

dism and or the wimserv files. If anything bricks i dont dick around anymore, clean install.

I had to do a rescue temporary install today, the main partition bricked. I was oop n running within 30 minutes. I will do the full proper reinstall tomorrow. And my profanity list didnt even get a single mention let alone a full and thorough workout.

 

[pmikep]

I could copy and paste all the dism and wimserv files from my last working NTLite'd Win7 iso to my current Win7. (Although I'm not sure which go where, with all the sysWOW stuff in Win7 but the raw dism files in "sources" on the iso.) But all my files dates look believable on my running Win7. That is, none of my files has a newer time stamp than February.

I deleted the dism log just in case it was goofing anything up. (Don't see how it could.)

I've got some work and deadlines to meet this week. So restoring an image from a few weeks ago will have to wait.

I still don't understand why that will fix it. But am willing to try.

 

[Clanger]

I dont even know if that would fix it, im grabbing at straws before suggesting going to the last resort of a full reinstall and seeing as you run raid will take time.

I dont know much about raid apart from mirrored or striped. You can noogy me now because i have just given another possibility. Data corruption on 1 or both drives "may" be the issue. When did you last run checkdisc? Have you run a hdd test tool like HardDisk Sentinel to check for the drives health? I have never had this problem this bad, i dont run raid. You never know,,,,,,

If you have to do a full reinstall, i would kill the discss first, write zeros to both of them, Active KillDisk Free will zero your drives. Install it to windows then create a bootable dos usb or dos (cd)iso, choose the dos option, its quicker than the linux option. It takes around an hour to completely zero a 500gb 7200 rpm hdd. Install windows then run checkdisc then install your drivers and programs etc.

If i get a totally hosed system, i write zeros, install windows, run checkdisc then install me drivers n programs. You can(but dont have to) run HardDisk Sentinel from time to time just to see how the drives are doing.

 

[pmikep]

I'm guessing that this is a reparse point problem.

If so, I found a cmd line to delete them. Example:
fsutil reparsepoint delete c:\server
A question for nuhi - what directory should I delete to fix my problem?

 

[nuhi]

pmikep, sheesh, sounds like the Windows mounting engine got confused and locks your image and/or mount directory.
The part that falls on me is the deletion being stuck, that I would like to fix, so please don't reinstall or format anything for now.
It's not as a big of a deal as it looks, those are 99% just NTFS links, so it's not taking much space.
That said, let's solve this.

- So you tried unmounting all of the mounted images on the Source page and clicking Forget on any of the images from the Image history list?
If so, did it pass deleting temp files or it gets stuck there even now?

- Did you change the Temp folder in the tool's Settings from when it was first mounted?
If not, please don't. Important is that the temp folder is the root of the mounted directory before trying the first step.

- It would not hurt to "chkdsl /f x:" your partition (X) containing the image and the one containing the mount directory, just in case.
From the elevated CMD.
Then repeat the step 1 if the deletion was stuck before.

- Then lastly, if you can get it stuck on deleting files repeatably, send me a %temp%\ntlite.log and I'll send you a temporary test version so we see more details.

Thanks!

 

[pmikep]

Hi nuhi:

Thanks for the quick reply.

I might make a video of what's going on, so that you can see what's happening for yourself, since the facts have been coming fast - and sometimes inconsistent, as reported by me.

In the meantime, I did run chkdsk /f on my C: drive (where my OS is). That didn't find any problems. But now I am running chkdsk on my S: drive - even though I have since moved my NTLIte Mount directory to my L: drive.

Now chkdsk on S is reporting all sorts of problems. "Inserting an index entry with id xxxx into index $SII of file 9."

And now it's "Repairing the security file record segment" and appears stalled. (Although how much of this is due to running unlocker or take ownership, I don't know.)

The hard drive light is on solid - I'll let it run for a few hours like this to see if chkdsk ever completes.

I wonder - if I make an image of my S drive (using Image for Windows), will that image clean out the reparse points/errors? If so, then I could image my S and restore, as a way of laundering the drive.

 

[pmikep]

Okay, I solved the problem. It took two actions: 1) nuhi sent me a beta of 1.7, which has a Purge feature for the mount and scratch Directories. 2) I extracted the CIRCoInst.dll from an old Win7 image (before I force closed NTLite and caused this problem) and I used it to replace the current one that seemed to be locking up my system, per my last screen shot.