Deployment Solution

I've built a universal image that I can deploy to several models of machines (Dell desktops, laptops, IBM laptops, etc.) without issue. However, a small subset of models fail to complete the distribute image job (using rdeploy). The rdeploy log shows the following error for all machines:

Error description:
The file path is invalid.
(Note: Current volume is NTFS from source slot 1.)

Possible causes:
A file that was expected to exist does not exist.

Possible resolutions:
Make sure the file exists and is accessible.
If the missing file is part of this program, reinstall the program.

==================== Technical details ====================
Logfile = created Thu Jun 04 11:11:02 2009
Build = RdeployT.exe 6.9 (9020)
Cmdline = G:\RDeploy\Windows\RdeployT.exe -noprompt -md -fh:\images\XPMaster.img -nooem -mcint:172.16.154.93

Status = 92 (0x5c)
Source file = imglib\fs\ntfs\ntfs.cpp
Line number = 999 (0x3e7)
Stack trace = 0x493096 0x49707d 0x41d61e 0x413348 0x413c5d 0x413897 0x414185 0x4149ef 0x40352a 0x412294 0x401db0 0x40e42d &Known=0x433e40
Path = /./WINDOWS/inf/iaahci.cat
Note = Current volume is NTFS from source slot 1.
Imaging library revision = 9020 (win32-x86-release build, Sat Oct 11 00:45:03 2008)

I noticed that the failing machines have OEM partitions (Dells) or as yet unused partition (VMs).  In both cases, the imaging task gets to 99% and fails with the error text above. Making this more interesting, if I deploy a different base image to any of these machines and then deploy the universal image on top of it, the universal image deploys without an error.

I opened a call for this issue when I only noticed on on VMs and the workaround was to use a ghost image instead of using rdeploy. Strangely, this worked. However, this is now affecting new models of incoming machines and I'm hoping to find an actual fix.

I've also gathered one last piece of information. After a failed imaging attempt, I tried to copy a file to prod: using firm copy. I also received an error there:

Error description:
The file seek operation failed.

Possible causes:
Data has been corrupted (e.g. in an image or on the network).

Possible resolutions:
Repeat the operation to see if the problem occurs in the same manner again.
Simplify the network by removing routers, etc. that might corrupt data.
If a file is being read, recreate the file.

==================== Technical details ====================

Logfile = created Thu Jun 04 11:23:27 2009
Build = firmpe.exe 6.9 (9020)
Cmdline = G:\firmpe.exe copy aclient.inp prod:aclient.inp

Status = 101 (0x65)
Source file = imglib\fs\ntfs\ntfile.cpp
Line number = 1574 (0x626)
Stack trace = 0x4324b6 0x42ca68 0x430e89 0x424588 0x42413e 0x40f55e 0x40f3f1 0x404067 0x405352 0x409a91 0x4010ef &Known=0x46d200
Imaging library revision = 9020 (win32-x86-release build, Sat Oct 11 00:45:03 2008)

Per Altiris documentation, 101 indicates that the drive is full. Ig I do a firm drives, I see that the prod partition exists and that it is over 150GB in size.

If anyone as any insight, I'm all ears.

Thanks!

Here's an off the wall question -Did you by any chance use image explorer to add/exclude files in your image? Sometimes the image index can fall out of sync with file data.

If so, try recreating the image without touching it with image explorer.

Kind Regards,
Ian./

Ian,

I did, in fact, use Image Explorer to edit the image. I'd considered recapturing it, but couldn't really find sufficient justification to motivate myself to do it. Great suggestion!

Though, I am skeptical about this because the image deploys to other models without any issues.

Thanks,
Mike

But...... i'm fairly confident you'll get a good result on this one ;-)

I now never edit images with ImageExplorer. Any file updates I make are done using automation scripts, and then later these are rolled up by creating a new image. Development are aware of this issue, but there is (to my knowledge) no KB to subscribe to, or a timescale set for a fix.

Kind Regards,
Ian./

Ian,

After some more exploration using firm drives, I noticed that the image copy stopped after c:\boot.ini. I opened up the image and saw that the next folder that it should've copied down was, in fact, excluded.  Right after the dir command displayed boot.ini, it returned an invalid drive error message.

I can't figure out why the image deployed without error to at least 8 other models but could not deploy to these three specifically.  Furthermore, if I applied any other image to these models first and then deployed the image in question afterwards, the issue went away. Suffice to say, this was a very odd problem.

I ended up deploying a different image to a GX760 then applying my master image on top of it.  From there, I recaptured the same base image (remember, the excluded files and folders were not sent to the disk) and have since been able to deploy it without any errors.

Thanks for the help!
Mike