The Deployment Solution Server stops responding when a set of automated jobs kicks off at their scheduled hour. It was determined that a large amount of these jobs were PC Transplant migration tasks.
How many PC Transplant Solution tasks can I run through Deployment Solution at the same time?
Altiris recommends that no more than 10 PC Transplant migration tasks are run at a time. In your environment, this could be even lower. This is not limitation of either PC Transplant or Deployment Solution, but a limitation of network and server resources. The following example illustrates this:
There is a 2 GB connection on the server, 100 MB on the clients. Without any contention or dropped packets, 100 MB connection *20 = 2 GB. This is the typical network configuration and when the server has more bandwidth on the network than 2 GB the bottle neck increases only to about 21 max computers based off a sustained throughput of 133 Mbit per second on each channel of a SCSI array. You can do the detailed math, but this is just where the limit of most server hardware will fall. If they have a 8 GB ATM to the server, and a SCSI array with a separate channel for each drive with 15 drives in the array and everything is running perfectly, if I recall correctly this would still have a limit of about 75 PCs. Since I have never seen a SCSI controller with 15 individual data channels for each drive, and 8 GB ATM is not typically available this is a nice theoretical limit, but not practical. The practical limit is 20 clients, with potential problems starting at 66 percent of that as TCP/IP has a 33 percent packet loss in most environments, thus problems could start to occur at 14 computers. If your network is more on the chatty side, and somewhere in-between for most environments, this number will be lower.
Using a PC Transplant logging level higher than 1 (default setting) increases the amount of data sent to the Deployment Solution and thus decreases the available bandwidth even further. Unless you are troubleshooting a specific issue, there is no need to increase the logging level. We have seen instances where setting error level to three results in even a single PCT job run through Deployment Solution to cause timeouts because of the extra data being sent through AClient to the Deployment Solution.