[Orca-users] Sol10 Patch Segment Fault

[Alex Kiernan]

On 24/04/06, orca at bias.org <orca at bias.org> wrote:
> Hello,
>
> I recently patched my systems and just noticed that orcallator stopped
> running (Solaris 10, Patch 04/2006, V240).
>
> This isn't critical to get working right away and I haven't spend a lot of
> time looking, but would like to know if anyone has seen the same
> problem/issues.
>
> Here is the tail of debug run trying to start the orcallator process:
>
> switch(i<27>)
> if_bge.number$ = number$<0> - (if_max[26]<-1> + <1>)
> refresh$(if_bge(STRUCTURE))
> name$ = if_bge.name$<bge0/0>
> ipackets = if_bge.ipackets<711149723>
> ierrors = if_bge.ierrors<429496729600>
> opackets = if_bge.opackets<1374567881>
> oerrors = if_bge.oerrors<0>
> collisions = if_bge.collisions<0>
> nocanput = <0>
> defer = if_bge.defer_xmts<0>
> nocarrier = if_bge.carrier_errors<0>
> ooctets = if_bge.obytes64<2049819451534>
> ioctets = if_bge.rbytes64<61413117979>
> ifspeed = if_bge.ifspeed<1000000000>
> iftype = <1>
> break;
> if (opackets<1374567881> != <0>)
> collision_rate = (collisions<0> * <100>) / opackets<1374567881>
> switch(ifspeed<1000000000>)
> break;
> i = <0>
> if (interface.number$<0> == pnetGLOB_net_size<24>)
> pnetGLOB_save_et[0] = gethrtime() * <1e-09>
> pnetGLOB_save_nets[0] = interface(STRUCTURE)
> net_count++;
> i++;
> interface.number$ = i<1>
> refresh$(interface(STRUCTURE))
> Segmentation Fault
>

If this is the same thing I ran across, the problem is that the additional kstat
structures Solaris 10 is generating, se doesn't understand, so doesn't
make the MAX_IF variable big enough, for the number of "interfaces"
netif.se pulls out. Basically some surgery needs doing on se itself.

I did make some proposals a while back, but I've not had any time to
actually go and look at the code.

-- 
Alex Kiernan