Hi, Michael wrote:
Hmm, I have very recent -current on two SS20s, one with 2x SM81, no problems there, easily survived building its own kernel. The other one has two 150MHz hypersparcs, which did cause problems in the past. It had two 125MHz hypersparcs before, which was mostly stable. There have been weird issues with hypersparcs before.
Indeed... there must be some issues with HyperSparcs. The old ones I had before they fried were a single-module and much slower. But I did run Solaris for many years, so I don't recall NetBSD status too well, but I remember stable, but it was also years ago!
The 150s are quite different: cpu0 at mainbus0: mid 8: Ross,RT626 @ 150 MHz, on-chip FPU cpu0: 512K byte write-back, 32 bytes/line, sw flush: cache enabled cpu1 at mainbus0: mid 10: Ross,RT626 @ 150 MHz, on-chip FPU cpu1: 512K byte write-back, 32 bytes/line, sw flush: cache enabled
Yours appear to be not just faster, but also a later revision.I got mine with great difficulty, not much choice finding a pair of HyperSparcs!
I tried again yesterday the MP kernel and I got "Watchdog Reset" after half an hour or so.
I wanted to test an idle machine.. but have disk issues now.
Seems stale so far, been updating kernel and userland with no issues. Will torture it a bit more once that's all done.
In case it works for you and not for me, maybe we should debug. Or should I try and install 9.4 instead and see how it fares?9.3 was unstable, but not to the point of crashing, If you remember it worked and after a a couple of days it would exhaust some strange memory nobody understood and I could do a reboot cleanly.
Riccardo