ProSphere and Solaris SSH discovery

I've recently been involved in an SRM suite deployment at our site, which is mainly an Oracle Solaris SPARC shop. I've noticed ProSphere's SSH discovery process with Solaris hosts has a significant flaw. When you first create and then run an SSH discovery job, ProSphere pushes its utilities to /tmp/nl_dwd on the Solaris host. Solaris, by default, has /tmp configured to use swap (volatile RAM). So when the server reboots, or somebody deletes the /tmp/nl_dwd directory, all subsequent discovery jobs run, but do nothing on the host, collect no data and ProSphere does not push the utilities to the host again. The only workaround I have found is to delete the discovery job and re-create it within ProSphere, so that it pushes the utilities again. Obviously this is not viable in a large environment. My questions are: Is EMC aware of this issue? If so, when can we expect a fix? Is there a better workaround? Thanks, Paul.