Tuning Question

[Scott Seekamp]

This week I started seeing this in my syslog:

kernel: Approaching the limit on PV entries, consider increasing either 
the vm.pmap.shpgperproc or the vm.pmap.pv_entry_max sysctl.

I changed the shpgperproc to 800 from 200, but I am still getting the 
message. These are my current settings:

vm.pmap.pmap_collect_active: 0
vm.pmap.pmap_collect_inactive: 0
vm.pmap.pv_entry_spare: 21690
vm.pmap.pv_entry_allocs: 12261136989
vm.pmap.pv_entry_frees: 12257750799
vm.pmap.pc_chunk_tryfail: 0
vm.pmap.pc_chunk_frees: 69895925
vm.pmap.pc_chunk_allocs: 69916210
vm.pmap.pc_chunk_count: 20285
vm.pmap.pv_entry_count: 3386190
vm.pmap.shpgperproc: 800
vm.pmap.pv_entry_max: 6447909

I didn't want to arbitrarily adjust the settings without understanding 
them. I could not find any good information out there on what to adjust 
these values to, or what to base them off of.

I recently added mod_perl2 to our lightly-loaded server so I'm guessing 
that is where this came from. Anyone have good information on how to 
handle this? The server is a 7.1-amd64 with 6GB RAM, load averages: 
0.20, 0.21, 0.26.

Thanks

Scott