DS 7.1 - "Reboot to Production" job after imaging reboots to Automation intermittently

[Migration User]

We've just begun to successfully deploy Windows 7 ghost images in a production environment. (This in itself is a cause for celebration!) We have the standard 3-task job of

1. Reboot to PXE (God I hate this task)

2. Deploy Image

3. Reboot to Production

We're using multicasting, and we've had to play around with various multicasting thresholds and time limits to get multicasting jobs to not time out. (For some reason, our DS server bogs down sending WinPE environments after about 10 computers, after which each one loads sloowwwwwly. Okay, regardless... we've tweaked the multicast settings to get them to work in our environment.

What I'm finding is happening is that the Reboot to Production job is not acting as it should on half or more of the imaged computers. Oh, it reboots alright - but it reboots them back into Automation, and they of course stay stuck that way until I send another Reboot to Production task. So, what gives here? How is it that I can send the same job to 20 computers, and 50-70% will process a "Reboot to" job one way, and the other 40% will process the job another?

Any ideas? We're seeing this with every Deploy Image job we run. Thanks

[Migration User]

Dustin,

once the "Reboot to Production" is executed, PXE gets a notice to modify itself and not to serve the client. the notice is acknowledged by a .sbs file with the following data:

Status=UpdateComputerStatus

MAC=005056C00008 (this will match your computer)

UUID={D5D1AF24-483F-11DB-BBDA-71824E800018} (this will match your computer)

SerialNum=

Image=""

SBSIPList=

 

 

so before the "Reboot to Production" happens browse to C:\Program Files\Altiris\Altiris Agent\Agents\Deployment\Task Handler\SBS\SBSStore  you could delete all those .sbs files they are just receipts.

[Migration User]

Nelo,

I tried that the last time I had problems, and it did resolve my immediate issue, but it's not really feasible to do every single time I image a batch of computers. How do I prevent this problem from occurring? Am I going to need to add a script into my imaging process to delete these files automatically? That shouldn't be necessary, imo.

[Migration User]

So don't even begin thinking along those lines.  Those files in the SBS Store folder are purely historical in value and are already processed, so removing and/or leaving them makes no difference at all.

The problem is a timing issue most likely.  Most likely, the system being rebooted is getting there before the SBS file is received by the SBS server.  We used to call this a "race to reboot" condition.  We don't see it anywhere near as often as we did in the past.

I would actually recommend you call this in so we can troubleshoot it and possibly log something against this.

We are, by the way, working on a permenant fix for this by modifying the way the PXE server is updated.  One suggestion, for instance, is to make a request in real-time for what is needed with a client rather than relying on a file-transfer methodology that might get delayed for any number of reasons.  We know this CAN happen, and that's why we're looking at alternatives.

Still, if it's moderately reproducable what you're seeing, we should capture this.

[Migration User]

In light of what Thomas just said, I guess the perceived "resolution" I saw from doing this was due to the restarting of the SBS Services? Though that doesn't make a lot of sense to me, either, since won't that make all computers show up as unknown briefly? I guess sometimes we just see what we want to see, facts be damned!

[Migration User]

Restarting PXE Server will clear the list, so that they take the "dafault" path for known computers - boot to production.

Keep your eye on the Beta listed in this forum - we should be fixing several of these "kinds" of issues in that release.  We hope...

[Migration User]

Thomas,

Sorry to be a pain - but if restarting SBS services clears the list so that clients take the default path for known computers, why do my techs complain to me every time I restart SBS services that all computers are acting as unknown until they fully boot into Windows? I.e., the clients present the "Press F8" prompt to boot into PXE. If no key is pressed, they boot to Windows. Known computers shouldn't do that, right? FYI, I I have "Respond to unknown computers" turned off. Sorry, I just do not understand this behavior. Thanks

[Migration User]

Hi,

We've recently started rolling out Windows 7 x64 much the same way as described in here and I also see random reboots into automation. They seem to occur alot more often after restarting SBS services.

This occurs on machines that have been booted into production several times and then all of a sudden boot back into automation.

It really is starting to be tiresome to experience how you screwed up the entire PXE process from DS 6.9.

I hope there are some fixes coming out for this soon. We are now on 7.1 SP2 it should be time for most bugs to be sorted out soon don't you think?

[Migration User]

To be fair, I don't think Deployment Solution 6.x started out great out of the box, but it is pretty solid now after some patches and updates.  Think we should see what fixes come out with the next release to pass judgement. 

The problems noted here I also ran into, but it's in my proof-of-concept phase and not production.  Production environment is still DS 6.9 untill the major bugs have been worked out.  Symantec guys have been good in posting up info.

What I've done, is have a 'reboot to production' task after deploying the image, then most of the time it doesn't boot into production.

 

[Migration User]

I'm not seeing this in the environments we've implemented.  We implement sites and subnets.  Have you implemented sites and subnets properly, where every subnet is assigned to a site, and each site has a site server?  I have seen this when people have not set up sites and subnets, such as early in the testing or pilot phase.

[Migration User]

Well we have sites and subnets defined. Even had Symantec partners in and verify the setup.

I believe it has something to do with the "answer pre-defined computers" in the PXE settings. We have this enabled since we have around 350 new computers that we want to deploy that are pre-defined.

I get the feeling that after restarting the SBS Services it defaults back to answer computers that were once upon time defined as pre-defined computers.

My question is when do computers stop being pre-defined? Is there a certain task I can trigger at the end of our deployment or something like that?

This happens sometimes days after we deploy the new computer and it works fine for several restarts then if we for some reason restart SBSservices or do other changes or just random it boots back into PXE automation.

[Migration User]

Here's a good read--which I can't compare right now to your environment--that might shed light:
http://www.symantec.com/docs/HOWTO21727

Known computers should never boot to Initial Deployment, regardless of what you do with server restarts.

Do you use automation folders?

Edit: You might prefer this diagram, which is more current than the above KB:
http://www.symantec.com/docs/HOWTO53220