Video Screencast Help
Endpoint Management Community Blog

Altiris SMP 7.1 upgrade from 6.x

Created: 13 Dec 2012 • 10 comments
Tim.Jing's picture
+4 4 Votes
Login to vote

This will be to document our migration from Symantec Client Managment 6 to SMP CMP 7.1. It is scheduled to be set in motion sometime Jan 2013 with the help of Xcend consulting services which also implemented our 6.x environment.

Our environment:

We are a school district with 80+ remote sites. 25,000+ Windows XP machines, a few Win 7, 300+ Mac's, 6,000+ iPads. We have a centralized datacenter, with fiber connections to each site. We have Cisco networking gear with gigbit connections at most ports, along with Cisco Aironet access points at some campuses.

We use Altiris for all new and existing SO and software automation except for the Apple equipment. The suites used in Altiris are:

Depoyment solution, we have 10 DS 6.9 SP2 servers divided up amoung the sites.

Inventory via NS, we have 3 client facing NS with a report server. Lots of custom reports for compliance and inventory questions.

Patch managment, for desktop and servers.

Software Delivery, for mass deployments.

Application Metering, for usage.

 

Observations prior to starting:

1. Symantec SE's are very vague about specs when it comes to environmental design. And since the platform specs changes often, it's even tougher to get an assuring answer on what physical hardware is required to put it all together.

2. Training is very tough to schedule as the low attendence of the courses results in them being canceled more often than not.

3. The Sale Engineers are very good at communicating and connceting you to the right party for any questions. But getting the right answers is another question.

4. The Support phone # is very inconsistent on who and what type of service you get. Over sea's calls tend to be very time consuming and takes some time for someone to get back to you with an answer.

5. The new console is bad, seriously more involved to do simpler tasks like creating a packaging.

So hopefully things will start well in the new year, unless 7.5 releases early and throws a wrench in our operation.

 

 

Comments 10 CommentsJump to latest comment

bmagurn's picture

I am in almost exactly the same situation as you, except I have been through an upgrade and we have so many issues with it, we have not gone live almost 8 months after consulting services.

 

I am in  a school district with 82 schools, 27,000 XP workstations & a handful of win 7. 

 

The new helpdesk is not the same product at all, it can be a powerful solution if you tweak the code under the hood, which they're supposedly removing the ability for us to do in the new version. 

 

The consultants I had didn't know how to setup the hierarchy properly and it took 4 months of calls to tech support and the tech support (Utah) guy bringing in someone from the other "product group" for Asset Managment to figure out that locations can't replicate. 

 

I'll message you my contact info so we can talk about issues, etc.

 

+1
Login to vote
Jtester's picture

We are school district of 160 schools 36,000 PC's.  Our infrastructure is 165 DS 6.9 SP3 servers.  When we were 6.9 for the NS structure we had 4 NS servers with 1 reporting server.  By the time we got all of our clients in there they had become virtually unusable.

So we had symantec come out and help us design the 7.1 system.  We ended up with a heirarchial system.  We have 1 parent and 2 children.  The parent has a site server and the 2 children each have 2 site servers.  Then every DS which resides at the schools has also become a site server.  We have a very beefy single purpose SQL Server. We had a consultant come out who was completely worthless.  I won't give his name because I have been told that he is no longer in that position.  

At this point about he only thing we have done with it is Micorosft Patching.  We are just starting 3rd party patching.  

The thing is that all of the schools love their DS servers.  I am very nervous about moving them to the new 7.1 system simply because it is so much slower and clunky.  Plus the secuity is so much more complicated.  I really only want a site tech to see their computers.

I will be interested in following your progress.

 

-5
Login to vote
Gibson99's picture

Tim, once you get used to the 7.1 console, you'll feel like you're going back to MSIE4 when you have to go back to the NS6 console.  and yet, there are parts of the 7.1 console that are virtually unchanged from 6.x.  i'm not saying it's perfect, but it is a big step up in a lot of areas.  oh and btw - parts of the console are still a little flaky in IE9 so keep IE8 around if you can.  dunno if that's fixed in sp2 (which i'd assume you'd start with, rather than the 7.1 sp1 we're on right now) but it's worth mentioning.

good luck.  i wish you infinite patience in dealing with the imaging part - we couldn't get it to work the way we needed it (18 package/pxe servers, one DS, one NS), and ended up keeping DS 6.9 for imaging/initial builds.  

our environment may be smaller, but the build process here is nearly the same as what you have at the district.  we still only have one NS 7.1 server and one DS 6.9 (plus ImageInvoker 0.4 - search connect, you might like it) and since we drop the SMP agent on at the end of the DS build job, it's not that hard to manage.  we just make sure to publish packages in ns7.1 on a public share on each package server so that the DS jobs can get to them to build the machines. 

you know how to reach me if you have questions or if you want to borrow any of my books from training.

If a Connect post helped you out, be sure to click "Mark As Solution" or the "Thumbs Up" button to let other users know about it.

-5
Login to vote
Tim.Jing's picture

jtester, you have a pretty crazy environment!

 

I'm going to try to update this blog everytime we make a stride or encounters a road block. But as of now, it's barely heating up.

 

So here is the current suggested design and I'm not really understanding the SQL Instance. We have 2 calls this after noon to sort it out with Symantec and Xcend.

 

 

-1
Login to vote
dodgint's picture

I would think the mulitple instances in SQL is to get better performance, since there would be a seperate tempdb and logs for each instance. As long as the tempdb and logs are on seperate LUNs for each instance.

Todd Clark County IS Vancouver,WA

+3
Login to vote
Ray OConnor's picture

Any update on this? I'd be really interested in knowing how your migration went and what resources you used.

Thanks!

Ray

+1
Login to vote
Tim.Jing's picture

Right click and save the picture if it seem too small.

+3
Login to vote
Jtester's picture

 

Timanator,
 
The picture that they have given you could be a copy and paste from the design that they gave us.  Here is where we ended with hardware.
 
Server Type   OS Processor CPU's Cores RAM
Site Servers Virtual 2008 R2 Std 64bit w/SP1 X7560 @ 2.26 GHz 2 2 8 GB
Children Virtual 2008 R2 Std 64bit w/SP1 X7560 @ 2.26 GHz 2 2 12 GB
Parent Physical 2008 R2 Std 64bit w/SP1 X5690 @3.47 GHz 2 24 24 GB
SQL Physical 2008 R2 Ent 64bit w/SP1 L7555 @1.87 GHz 4 64 64 GB

 

-3
Login to vote
Tim.Jing's picture

Hey Gibson99, Cant believe you guys are still using it after all the talks of SCCM? 

Agree the console, it's not a very good/consistent design. There are areas where the "save" button needs to be pressed. Some When you click "OK" or "Apply" the page does not close, and you need to hunt and peck and scroll to verify there isnt a "Save" option.

 

Also from a glance, it looks like there is more steps just to create Jobs/Tasks. It's like this was designed by an accountant, who wanted every step to have it's own folder. And then you can daisy chain everything instead of creating one job for something simple. Disc imaging is the perfect example of multiple steps to accomplish something simple.

 

Jtester,

The Symantec SMP manager thinks we can go with this design as long as we can provide the IOPS. Which is daunting even with the 3Par San. 

With Each SQL needing about 2,000 IOPS, with 1 SQL server and 2 instances, this is easily accomplished with the 15k discs on a SAN. However, whether the initial IOPS in the design is enough remains to be seen. 

My math are 130 IOPS for each drive, multiply by 49 drives is about 6,370 IOPS total not counting RAID cost. But thats also being very conservative with the drive campabilities.

 

 

-3
Login to vote
Gibson99's picture

nothing in enterprise moves that fast.  besides - weren't we just starting to look at 7.1 when i left the district?  that's been over 2 years now.  

i'm just glad we don't have to deal with the scale of what you guys do.  we only have about 2000 clients worldwide.  it's the ones in china and the middle east, with crappy WAN connections, that cause us any trouble.  and those offices are so small (30-50 people each) that it's not worth setting up even a secondary NS for them.  

If a Connect post helped you out, be sure to click "Mark As Solution" or the "Thumbs Up" button to let other users know about it.

+1
Login to vote