Problem regarding imaging and netbios name


Recommended Posts

hello gang,

i have a new problem which seems to have stumped me. I recently updated an already created image (Ghost), created a new image of that system then rolled it out to approx 30 client computers. the procedure i use is this:

1 Install O/S and apps

2 Configure (using a specifically created admin account) profile to copy to default user

3 run sysprep (with the option left on to regenerate sids)

4 and whilst the computer is DOWN (as in before it boots to create new sids and follow my sysprep.inf answer file) i ghost that image.

So theorectically i can rollout fresh images that havent finshed initializing or joined domains yet, to ensure these are REALLY individual images. sysprep then continues and domain attachment is automated via sysprep.inf so i end up with computers ready to rock and roll.

My problem now stems from the fact that on my LAST set of images, the netbios name seems to have been copied across to every image. Now this is peculiar as when i look at each computer in the my computer properties, computer name, more, netbios name, it actually says the computer unique name. As expected ofcourse. But when i ping a recently imaged computer from another computer with the command "ping -a NEWIMAGE1" i recieve a reply from the correct ip adress, but of the netbios name of the host image computer before sysprep was run on it.

Let me say that bit again, the host image (the computer my image was created from) had a computer name of 'Image010306Builda', and such a netbios name of the same up 2 15 characters ofcourse. Sysprep was run on this computer to regerate sids and the like. When this computer shutdown after sysprep i imaged it. My new images all got new computer names (during the sysprep procedure). But now when i interogate these computers they have the same netbios name as the originaal host computer, whilst their own netbios name configuration IS their own (correct) dns computer name.

i cant seem to find the cause of this. I had a WINs server in operation but removed it for clarity sake, it did not have any records regarding my issue anyway.

Any help please, especially from S.O.E admins or anyone really. Thanks alot guys.

Hi amfony,

This may sound strange but we had exactly the same issue at our DR site recently.

You'll find all the machines are pinging via WINS instead of DNS i.e. WKS001 instead of WKS001.domain.co.uk)

Anyway, the fix the engineering team implemented was to script in an "ipconfig /release" and "ipconfig /release" upon the final boot (after our image was rolled and had at least once booted into windows).

Try that on one machine and then try pinging...you should be able to via DNS and resolve name correctly.

What I couldn't understand was a reboot was NOT resolving the issue...it required a forced ipconfig release and renew. Very strange indeed.

Please post an update here whatever the outcome.

Many thanks

i tried but no dice. under every circumstance i get the same reply. i even ran sysprep again on a image client machine, removed ut from domain, uninstalled the nic, rebooted, searched registry, i disabled netbios expliocilty from the nic, (its disabled by dhcp by default) and no dice. i am at a loss

Where are you resolving these names from? Your machine? Every other machine on the network? Where?

Sounds like you have a cached entry or lmhost/host entry in where ever your trying to access these machines from.. What does a nbtstat -A IPADDRESS of the machine in question show?

What does your cache show on the machine your doing these queries from? nbtstat -c

What does the machine in question show for nbtstat -n show?

Sorry for the late reply guys! but i am back now

Sorry budman i should have been more specific, from the newly created image (lets name it CLTIMG) when i ping it self, i get returned its proper "dns" fqdn (and proper address), when i type the command "ping -a IPADDRESS OF SELF" i get the CORRECT DNS name relpy again. Now this is what i expect as i have disabled netbios via DHCP options.

When i ping from ANOTHER (any other) computer to the CLTIMG via its proper name, i get a correct and legitmate dns response (my ddns is working fine and all records are correct for the new images), from anyother computer when i type "ping -a IPADDRESSOFCLTIMG" i get resolved back "ORIGINALIMG" (being the image i copied to rollout to client machines) and the ip address of the CLTIMG box.

Sorry if i am not very clear, but when resolving from it self, the CLTIMG thinks everything is ok. When resolving netbios from another computer it gives me the wrong computername.

I did go through pretty much nbtstat had to aoffer me budman and to confuse me even further when i typed "nbtstat -a CLTIMGIPADDRESS" i get the CORRECT netbios name returned to me.

in the cache if the CLTIMG i get nothing the domain name (which is wierd as i shudnt have any entries here is my netbios is disabled) and nbtstat -n gives me once again my correct computer name.

Interesting no?

I am thinkning along the lines that sysprep itself didnt quite work as supposed to rather then any configuration errors. What is your line of thniking budman? and others ofcourse?

Thanks guys and sorry again about the late reply

I've built plenty of machines and joined them to a domain for tweaking before ghosting, BUT I always remove them from the domain and remove the comp account in AD prior to actually making the image. I havent seen any odd issues form doing so and usually use sysprep to have them prompt for a new name prior to joining the domain on first boot after applying the image...

The problem as it sounds to me is more of a DNS issue than a workstation issue... Maybe try removing the DNS entry for the particular machine that your pinging then renew the IP of that machine and have a new DNS record created. Then of course ping it from somethign else... OR, maybe remove the problem machine from the domain, remove the comp account in AD, rename it and rejoin the domain... Either of these will basically allow DHCP to update DNS with a new record for that machine, and hopefully resolve your problem.

ok that could be the problem, the fact that i DONT remove the computer from the domain b4 sysprep-ing. I was under the assumption that a reseal on sysprep was the equivalent of "making the system ready to be used by end users", and therefore assumed all naming and domain membership was revoke and nullified before sysprep was finshed its job. As well as re-sid-ing.

Now ill have to test this tomorrow, but just a thought. I too, have built plenty of computers, it seems in error with my mistake regarding the domain. But i have now in a prodcution environemnt 95% of my images running well. As in NOT WITH THIS PROBLEM. i dont feel that my procedure this time was any different from the other systems i have built then imaged.

I called micro$oft, with a capital SOFT, and i explained my issue, as soon as i mentioned "3rd party deployment application" they washed theirs hands of the problem and said "we cant guarentee anything with 3rd party products".

As for UnaBonger, the records i have delted and to no avail, the first post actually refers to new ip lease and that was no dice either.

Im back wit the results, I did create a new image with removing the domain, this didnt seem to fix it.

I have recoursed back to creating a brand new installation of XP from my custom install CD. Customised, removed from domain as suggested in this thread and imaged. This works.

Unfortunatley i cannot pin point the problem, but a work around will do until i get more time to investigate. Thanks for all your help guys with this issue.

Sincerley

Amfony

This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.
  • Posts

    • Yes, agree 100%!! It baffles me that so many people are so outspoken about hating Win11 but have taken zero steps towards trying out other operating systems. If you don't like the direction Microsoft is taking Windows, then staying on Windows 10 is only the short-term solution, clearly the long-term solution is to find an OS you prefer. Let's be honest about two things...Windows 10 was FAR from perfect, it makes many of the same offenses that 11 does, but maybe to a lesser degree. Also, if anyone is holding out hope that Windows 12 will be some kind of return to the good-old-days, they are being extremely naive. What will end up happening is they will get dragged along to W11 once W10 support fully ends (and they realize 10 IoT isn't what they thought it was), just like they got dragged from Windows 7 to 10. Then start the cycle over, they will figure out how to make W11 work the way they want it to and will start complaining about how bad W12 is.
    • only works for 1 microsoft account . im on rp channel and have 2 pcs with the same microsoft account, i pc got it the other one did not
    • Your first comment about one needing to send all their data to Microsoft's "cloud" to continue Windows 10 support was of course ridiculous but I can't disagree here. I am not sure why the OP would wait to switch to Linux if their Mom just uses a browser as she can use the same one with Linux. Why would she care what OS is running? Install /configure Linux Mint or Ubuntu LTS and almost no maintenance really.
    • Seagate Exos 20 TB BarraCuda 16 TB 7200 RPM CMR hard disk drives are limited-time deals by Sayan Sen Currently, there are some great sales on SSDs, in particular, NVMe M.2 ones, like this SK hynix P41 Platinum P41. However, if you are in the market for a hard disk drive (HDD), then Seagate has you covered. That is because two high-capacity drives, the Exos X20 20TB and the BarraCuda 16 TB are currently at great prices (purchase links down below). Since both of them are high capacity models, they are based on CMR or Conventional Magnetic Recording technology which is typically great for robust use cases like on NAS or (media) home servers and similar such scenarios. However, the Exos is enterprise-grade quality so it is definitely what you would want for 24x7 usage, but the BarraCuda can also serve pretty well but will be a better option for backup purposes. The Exos X20 promises power-on hours of 8760 per year, while the BarraCuda claims 2400. The latter comes with double the DRAM cache, though with 512 MB of it, while the former has 256 MB, but it should still be plenty. In terms of data transfer speed, Seagate promises up to 285 MB/s (272 MiB/s) from the Exos X20 and the BarraCuda is rated for up to 190 MB/s. These are the maximum sustained data transfer rates for the two drives. Since the Exos X20 is an enterprise-class product Seagate does disclose its random performance throughput, which is 168 IOPS for reads and 550 IOPS for writes, respectively. While the BarraCuda's randoms are not specified it is expected to be lower and to compensate, it packs double the DRAM cache. Get the HDDs at the links below: Seagate Exos X20 ST20000NM007D 20TB SATA 6.0Gb/s 3.5" Internal Hard Drive: $379.00 + $50 off with promo code SACET2372 => $329.00 (Newegg US) Seagate BarraCuda ST16000DM001 16TB SATA 6.0Gb/s 3.5" Internal Hard Drive Drive: $194.99 (Newegg US) This Amazon deal is US-specific and not available in other regions unless specified. If you don't like it or want to look at more options, check out the Amazon US deals page here. Get Prime (SNAP), Prime Video, Audible Plus or Kindle / Music Unlimited. Free for 30 days. As an Amazon Associate, we earn from qualifying purchases.
    • Why do you think the writing is on the wall? Tesla continues to have the highest selling models on the planet, and they continue to grow on that despite their junk reliability rating. I'm not going to buy one because of the continued build quality issues (and lack of Carplay), but I appreciate that they're the best player in an EV market that is massively lacking in infrastructure and affordability. Besides the EV technology needing to eventually outpace gasoline, self-driving technology is undoubtedly going to be the future. It's not going to happen tomorrow and not even in 10 years, but within our lifetimes, most cars on the road will be self-driving. That's the only way to solve the issue of a million deaths on the road. Waymo is using technology that's simply too expensive, which means it has to charge high fares. Tesla is going to blow past them with their ~$40k car and further yet when they release that weird looking 2-seater. Pretty soon Nvidia is going to be a market player, Amazon will be a player, and possibly the other 20+ startups that are working on a vehicle. I'd like to revisit this in a year and see how the market looks. I'm willing to bet there will be a huge increase of driverless cars on the road, and I doubt that Waymo is going to be the majority of them. Waymo isn't profitable for Google, so it's entirely possible the plug gets pulled as soon as several players leave them in the dust.
  • Recent Achievements

    • Week One Done
      DrRonSr earned a badge
      Week One Done
    • Week One Done
      Sharon dixon earned a badge
      Week One Done
    • Dedicated
      Parallax Abstraction earned a badge
      Dedicated
    • First Post
      956400 earned a badge
      First Post
    • Week One Done
      davidfegan earned a badge
      Week One Done
  • Popular Contributors

    1. 1
      +primortal
      616
    2. 2
      ATLien_0
      229
    3. 3
      Michael Scrip
      167
    4. 4
      +FloatingFatMan
      161
    5. 5
      Xenon
      142
  • Tell a friend

    Love Neowin? Tell a friend!