OS Deployment is EXTREMELY SLOW

[lt.son]

Hello all,

 

I hope I can lay this out clearly so someone can understand what's happening here and help out.

 

We've had OS Deployment working for a couple of years now with no real issues. We have about 30 offices nation wide. We have a single Site and a Distribution Point at each location. The DPs are PXE enabled. They are all virtualized except two, which are located IN ADDITION to the primary DP VMs in our two largest offices.

 

Both of these locations have multiple subnets. They both had speed issues on day one when we enabled OSD. After much troubleshooting and trying every suggestion we tried, we finally got "regular speed" by installing a physical DP on the subnet that the local IT staff were on.

 

Two weeks ago we upgraded the WAN circuits in the 2nd office. We replaced the router and all switches. Because this location now houses our disaster recovery site, we rejiggered the network and added a new subnet that is now where all the clients are located.

 

Now we have the same issue we had when we setup OSD. It took over 30 hours to image a single system. We have moved the physical DP to the new subnet, but PXE is not working now. I can get PXE working, so I'm really curious about the slow deployment across subnets in the same office.

 

Our setup:

HQ, 31 Subnets, virtual Site Server & DP, physical DP** on ITs subnet

31 remote offices with a single subnet and VM DP

1 remote office with 3 subnets a VM DP and a Physical DP** on client subnet

 

** Without 2nd DP imaging takes more than a day.

 

[anyweb]

start looking closely at the smsts.log to see where it's pulling the content from, it could be that your boundaries are incorrectly defined and therefore it's pulling content across the WAN (i.e. slow)

[YPCC]

as above, incorrect or no subnet at all (so might be pulling content from a fallback DP)

 

Failing this, test connectivity between location of imaging and DP. Perhaps QOS at network level is slowing things down.