3.81 - Connection Closed, Session Terminated ?

[Chris Stephens]

I just recently upgraded from RDP 3.50 to 3.81. Prior to upgrading, when PXE booting a client, it would load the Altiris Agent and say something like "awaiting instruction from server" and if a job was submitted it would pick it up and run with it. Now post-3.81, when I boot up a client, it says "Client record Updated" followed immediately by "Connection Closed, Session terminated" and it auto-reboots itself and just keeps going through this cycle.

I've run across this when a server's MAC is already present for a server under Computers but this one is not. Any other ideas?

[Chris Stephens]

Oh and I should mention that the server does not appear in the console.

[Chris Stephens]

OK actually this appears to be the new default when PXE booting a server that does not have an OS on 3.81. Previously on 3.50 it would PXE boot into the Altiris client and sit and wait for a job to appear for it. Now it boots and if there isn't a job waiting for it, the client exits. To me this is sort of half-assed as previously you'd boot the client just once to get it into the console and send a job. Now you have to boot it once to get it into the console, send a job to it and then boot it again in order for it to see the job. If this was done this way on purpose, it's definitely a step backwards.

[Gordon Leonard]

The process is

Boot and wait for a job.
Drop a job.
Server in "wait mode" runs job.

You should not have to reboot to run the dropped job.

[Chris Stephens]

Yeah I know that's how it used to work for me. I would boot a new server for the first time and it would appear in the Deployment Console and the client would sit and wait for a job to be submitted. Now on 3.81, I boot a new server and it appears in the console and the aclient will exit immediately with "Connection, Closed, Session Terminated." I then have to submit the job and PXE boot back into the aclient a second time for it to pick up the job. Per HP support, its SUPPOSED to work this way now which I find hard to believe. It's really half-assed on Altiris' part if they intended it to work this way as I now have to boot a new server twice to run a server build on 3.81 vs. once on 3.50. I figure there has to be something wrong or if it's intended to work this way, there may be a way to override it.

[Gordon Leonard]

What is your case number?

[Chris Stephens]

It was 3603539585 but I believe it is closed now.

[Gordon Leonard]

Have you been able to get this resolved via the HP Helpdesk?

[Chris Stephens]

Nope. They say it's supposed to work that way.

[Gordon Leonard]

Could you check to see if your lookup keys are set.

Tools -> Options -> Global [Tab]

(see attachment)

[Chris Stephens]

Yep, same options set as the picture, Serial Number and MAC Address.

[Gordon Leonard]

When your PXE boot for the first time (and RDP has never seen your server) does it PXE boot into the LinuxPXE?

Also you noted that you delete the server from the RDP console (because you want it to stop and wait) can you also confirm that the server is not visible even from the phyicial view. [View -> Show Phyicial Devices].

[Chris Stephens]

It PXE boots to a menu where I can choose WinPE 2.1 Managed or Linux Managed. When I remove the blade server from the Computer view, it is also removing it from the Physical view.

[Gordon Leonard]

Can you check you PXE configuration (Tools -> PXE Configuration)

Besides the IP address does it look like (see attached)

[Chris Stephens]

Yep, I've tried toggling between Linux Managed and WinPE 2.1 Managed to no avail.

[Gordon Leonard]

Make sure all the other items are set also (not the IP). The PXE boot should not be asking you to pick a PXE boot.

Something in RDP is not seeing a new server as a new server. If it's not in the console it's a new server and it should stop. Since you can drop a job during one of the many reboots I know your mapping a drive to the RDP server.

Your saying the server does not exist in the console. You turn on the server and it PXE boots and checks into the server, but instead of waiting it goes off and processes nothing. It's almost like you told RDP that when it sees a new server run this job of nothing. If you click on the server which just checked into the RDP console do you see and task/job that are running/ran against it?

[Chris Stephens]

Nope, no tasks are assigned to the new job. Something within the Altiris client must be telling it "hey there's nothing to do here, please exit."

[fluidville]

Getting the same problem, all seems to work fine on the 3.80 version, am running a test enviroment for 3.81 and, getting the same problem. with the client saying updated then going into session terminated. with no pending jobs.

[fluidville]

May be barking up the wrong tree here, but could this be something to do with the SQL, are you running and express version of SQL to support your altiris database.

[Chris Stephens]

Nope, we have the database off on a separate SQL 2005 server.

[fluidville]

Chris,

I have managed to resolve the problem our end, it came down to the licece we were using was not compatible with version 6.8. if you check your licence or licences you will probably find they are 5.5 versions, you will need to upgrade these to 6.8.

I believe the upgrage can be done by sending your existing licence to HP.

Hope this helps.

[Chris Stephens]

How do you tell what version the licenses are? I believe this was a new RDP 3.50 (Altiris 6.8) install before I took it over and upgraded to 3.81 (Altiris 6.9) so I know the licenses are probably 3.50 but can you see the version any where?

[fluidville]

Chris,

if you open up the console, got to Help then about, an about window will open, bottom right of this screen you will find a scroll window scroll down till you see.

Product: Altiris eXpress Deployment
Server 6.8

In your case you may have 5.5 version of the licence. Also another thing to note on this screen is how many nodes you have licenced, if your database has exceeded the alotted amount you will have the same problem you have been experiencing. meaning if you have 10 nodes licened you can only have 10 machines listed on your deployment console.

Hope this helps.

[Chris Stephens]

We eventually found out that our network team changed the DHCP relay/scope for the subnet that this server existed on around the same time we performed the upgrade to 3.81. The new scope was providing multiple gateways to the servers when they DHCP-booted, 3 of which were invalid and listed first. Once the DHCP scope was changed to include only the existing gateway, this problem went away. So now when a new server boots into the Altiris client, it sits and waits for a job submission like it is supposed to. I knew the session terminated was not designed this way in 3.81 as was stated above.

[Erick Soulages]

Hey,

I had the same problem and i solve it with regenerated the PXE boot packages in:

Tools -> PXE Configuration  ->  boot menu.

In image Boot Regeneration, Selected the two "Linux" "Winpe" and click Regenerated.

Bye,

Erick.