Solaris Common Messages and Troubleshooting Guide

Panic on cpu 0: valloc'd past tmpptes

Cause

The machine is an SS20 with 256 Mbytes of RAM, an FDDI interface, and a single CPU. It is running Online Disksuite for mirroring and striping. The following recommended kernel patches were installed:


102517-03 
102436-02 
102394-02 
102516-06
After their installation, the machine was rebuilt to allow for the new patches to be implemented. However, the machine panicked immediately after loading the kernel with this error message.

Action

The kernel was rebuilt with a new MAXUSERS value of 96, and this kernel enabled the machine to boot properly.

Technical Notes

Information directly related to this situation was not available; however, there was a description of another type of panic that was related to seg_u. In that description, the MAXUSERS value was set too large, causing the kernel to overrun table space. Furthermore, the value of MAXUSERS varies among the different architectures and the different revisions of the OS and is directly related to the amount of physical RAM in the system in an inverse proportion. Further investigation revealed that the value of MAXUSERS was set to 128. Based on the related information, it seems that the panic was due to valloc attempting to define memory space in excess of the value of tmpptes.