0

I frequently see that deployment execution on our development WAS server hangs, while the applications running on it keep on responding to user requests. This leads to deployments taking ages, with frequent reboots either of the applications being deployed or even of the whole application server.

Since this is a dev environment, the apps deployed don't need to be always available, so I was thinking of changing the process priority execution of either the node agent or dmgr, so that the deployment process may keep on running while a few requests may timeout/CPU starve.

However, I'm not entirely sure which one of the two components' priority to modify: from what I gathered the dmgr just manages the process, while the node agent does the grunt work, so in order to keep the deployment process to starve I should increase the node agent's priority, correct?

Orgrim
  • 101

1 Answers1

0

I think you have the relationship between the processes a little mixed up. The node agent is the simplest -- it keeps other processes running and provides basic remote control from the dmgr.

The dmgr and the individual application servers to deployment of one kind or another.

Unless you can observe a very high run queue or very high CPU usage, I think it's more likely that you are exhausting the dmgr's java heap and should tune that first.

covener
  • 1,665
  • 9
  • 15