On Wed, Oct 10, 2001 at 04:03:05PM +0200, Manuel Bouyer wrote:
> Hi, 
> I have a micovaxII 3300, booting diskless of another NetBSD machine.
> For some time I've not been able to get a working -current kernel.
> A kernel compiled 4 days ago hang once getty is started; login from
> network is impossible; the sshd process doesn't anserw.
> The machine is still anserws to ping, though.
> 
> A previous 1.5X kernel would come up multiuser but once a user logged
> in, any command would fail with 'no more processes' (or something like
> that, the kernel didn't have enouth resources to start a new process).

I've been able to get this with my 1.5Y kernel too (by starting less stuffs
from rc.d). 
I can log in on the console but a make depend in the kernel dir dies
with "Cannot fork", so it looks like the kernel doesn't have enouth
resources.

Here's the output of a vmstat -s:
4096 bytes per page
     2556 pages managed
      734 pages free
      750 pages active
       87 pages inactive
        0 pages paging
      450 pages wired
        1 zero pages
        0 reserve pagedaemon pages
        0 reserve kernel pages
        1 minimum free pages
        5 target free pages
      533 target inactive pages
      576 maximum wired pages
      146 swap devices
       64 swap pages
      499 swap pages in use
       25 swap allocations
       12 anons
       10 free anons
        5 total faults taken
       10 traps
        1 device interrupts
    16000 cpu context switches
      150 software interrupts
      120 system calls
       30 pagein requests
     2795 pageout requests
    18394 swap ins
    18394 swap outs
    17741 pages swapped in
    23302 pages swapped out
    34730 forks total
        0 forks blocked parent
    22148 forks shared address space with parent
        0 pagealloc zero wanted and avail
    48465 pagealloc zero wanted and not avail
       30 faults with no memory
        7 faults with no anons
        9 faults had to wait on pages
        0 faults found released page
      150 faults relock (245 ok)
       39 anon page faults
       47 anon retry faults
        0 amap copy faults
     5294 neighbour anon page faults
        0 neighbour object page faults
    13964 locked pager get faults
        0 unlocked pager get faults
        0 anon faults
        0 anon copy on write faults
        0 object faults
        0 promote copy faults
      688 promote zero fill faults
      688 times daemon wokeup
     7231 revolutions of the clock hand
       30 times daemon attempted swapout
     2435 pages freed by daemon
      120 pages scanned by daemon
     2080 anonymous pages scanned by daemon
    10799 object pages scanned by daemon
      658 pages reactivated
     4436 pages found busy by daemon
    10049 total pending pageouts
      750 pages deactivated
    22483 total name lookups
          cache hits (84% pos + 1% neg) system 0% per-process
          deletions 0%, falsehits 0%, toolong 0%


Any idea on how to solve this ?

--
Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
--