The following problem only applies to Solaris 2.6 systems. However, you can install patch 107308-12 to correct the problem on Solaris 2.6 SPARC Platform Edition systems. On Solaris 7 systems, it is fixed in CDE v1.3 build 19.
If the current scheduler's behavior is changed, it will uncover a race condition in the signal that is used to communicate between dtlogin and Xsun. This can prevent dtlogin from restarting after exiting a CDE session. To avoid this problem under Solaris Resource Manager, use dispadmin(1MSRM) to set a smaller time quantum.
Extract the current time quantum parameter:
# dispadmin -c SHR -g > /tmp/dispadmin.out |
/tmp/dispadmin.out should look like:
# # (SHR) SRM Scheduler Configuration # Resolution=1000 # Resolution Quantum=110 # Global time quantum for all processes |
Edit dispadmin.out and change Quantum to 50. Input the change:
# dispadmin -c SHR -s /tmp/dispadmin.out |