Deployment Solution

I have some imaging jobs where the first step is boot to PE

if I have a machine in the console and assign the job and it reboots things appear to work fine; however, when I have new machines and I use network boot then F8 and tell it to go into PE they don't seem to be as reliable.  I have the job menu configured and it will come up for new machines and if I select a job I would say maybe 90% of the time it shows that it is waiting for the pxe task to respond success.  The other 10 it does and the job continues as normal.  We used to see it work a lot more.

Someone mentioned to me that in PE sometimes the task server it will pick up will be the NS and when that happens this is typical.  So looking at the jobs that sit there and do nothing sure enough the task server it shows they are connected to is the NS, if I stop the job and simply restart it then it gets a different task server and works fine.

This is very frustrating for my techs who are trying to image machines as they do like the option of the menu when taking a machine out of the box vs having to use the console to wait for the machine to show up and then assign the job.  Is there any known fixs for this at all?

We are on 7.5 SP1 Hotfix 5.

Sonic,

We have been dealing with this since we upgraded to 7.5.1 HF5.

We see it happens randomly about 10-15 percent of the time.  It will show it hit the SMP task server instead of being assigned to the task server/site server in the machines own site.  Stopping it and then restarting the task makes it trigger.

It only does it when using Initial Deployment menus, which we specifically like in this new version because we can show it to both managed and unmanaged machines.

Previously, 7.1, I made a custom pop up for the service desk agents to get the proper computername that the client has in the console so they can locate it and assign it the imaging task.

Doing it that way in this version works 100% of the time but we are still using the initial deployment menu despite this.

We have been going back and forth with Symantec for a few months regarding this and even their backend DS guy cant figure t out yet.

They have suggested 2 things to try and remediate it:

1.  On the SMP, under IIS, under /Altiris/ClientTaskServer they had us use the IPAddress restrictions option.  We basically denied all IP's to hit this part of IIS except for the SMP's IP and localhost.

This was trying to thwart any other task agents from connecting to one of the 100 slots that the SMP can handle.

It did not work.

2.  Symantec tech had us remove the Task Server service from the SMP.

This was very unpredictable.  While it worked, in relation to WinPE tasks never being assigned to the SMP, it sometimes had to be added back and remove again to fully remove itself.  We then had to make sure we rebooted the SMP as well as all remote task servers or the remote tasks servers would fail to send up web statuses of executed jobs and just site saying Pending.

This though seemed to be only in vain.  The SMP and remote task servers would slowly show odd task/job issues from 1-3 weeks after removal which causes us to reinstall the Task Server back on the SMP.

This is where we are at.  Living with it for the time being and still have a ticket opened.

Hope this helps you and all others who have this oddity.

Thanks,

Clay