I had almost the same problem here too. Box is dell 1850 with MP-board, 1 CPU inside, running GENERIC.MP. It stopped to accept any connections, but: 1. responded to ping very quickly 2. syn-scan gave out that ports were open (those which should be open) 4. no heavy traffic incoming 3. no dumps on console were detected.
I had to switch to GENERIC. I'll see if it will hold up. Box is remotelocated and was previously running 3.7 without any problems. On Monday 10 July 2006 10:56, you wrote: > It happened again: one of our servers with OpenBSD i386 3.9-current MP > of a couple days ago crashed with the following kernel error: > > uvm_fault(0xd071cd68, 0xeb64c000, 0, 1) -> e > kernel: page fault trap, code=0 > Stopped at pmap_page_remove_86+0x114 movl 0(%eax,%edx,4),%eax > > Unfortunately I wasn't there, so there is no traceback... > > This is not the first time it happens since we: upgraded to 3.9 (then > -current), installed an X2 CPU, switched to MP kernel. > > From the above error message, can you say if this could be a kernel bug > or is more probably an hardware problem? > > > Thanks.
