[Orca-users] No output for Solaris 8

[Jeff Zacharias]

I have a problem with a single Solaris 8 box in our environment.  Only 
this single box does not produce any output into the 
/usr/local/orca/var/orca/orcallator/$HOSTNAME director, although the 
orcallator.pid file does get updated everytime orca starts.  We are 
running orca version 0.27b2, with Se toolkit version 3.3.  We have 
packaged up both orca and Se toolkit, and have installed these same 
packages in the rest of our environment, and no other box has these 
problems.  For debugging I have tried turning on the "-d" flag in the 
start_orcallator script.  Here is some of the output from the nohup.out


if (timestamp<0> == <0>)
timestamp = time(<0>)
pvm_update(timestamp<1052854494>)
if (timestamp<1052854494> == <0> || timestamp<1052854494> - 
GLOBAL_pvm_update_time<0> >= <2>)
GLOBAL_pvm_ncpus = sysconf(<15>)
for(i = <0>
; i<0> < GLOBAL_pvm_ncpus<4>; i++)
i = <0>
p_vmstat$GLOBAL_pvm.number$ = i<0>
GLOBAL_pvm[0] = p_vmstat$GLOBAL_pvm(STRUCTURE)
ncpus = sysconf(<15>)
if (initialized<0> == <0>)
pvmGLOB_cpu_size = ncpus<4>
pvmGLOB_etime = new double[pvmGLOB_cpu_size<4>]
pvmGLOB_old_time = new ulonglong[pvmGLOB_cpu_size<4>]
pvmGLOB_cpu_sysinfo = new ks_cpu_sysinfo[pvmGLOB_cpu_size<4>]
pvmGLOB_old_cpu_sysinfo = new ks_cpu_sysinfo[pvmGLOB_cpu_size<4>]
pvmGLOB_cpu_vminfo = new ks_cpu_vminfo[pvmGLOB_cpu_size<4>]
pvmGLOB_old_cpu_vminfo = new ks_cpu_vminfo[pvmGLOB_cpu_size<4>]
n = kstat$misc.clk_intr<0>
for(i = <0>
; i<0> < ncpus<4>; i++)
i = <0>
csi.number$ = i<0>
refresh$(csi(STRUCTURE))
pvmGLOB_old_cpu_sysinfo[0] = csi(STRUCTURE)
cvi.number$ = i<0>
refresh$(cvi(STRUCTURE))
pvmGLOB_old_cpu_vminfo[0] = cvi(STRUCTURE)
pvmGLOB_old_time[0] = n<682246>
pvmGLOB_etime[0] = n<682246> / hz<100>
i++;
csi.number$ = i<1>
refresh$(csi(STRUCTURE))
pvmGLOB_old_cpu_sysinfo[1] = csi(STRUCTURE)
cvi.number$ = i<1>
refresh$(cvi(STRUCTURE))
pvmGLOB_old_cpu_vminfo[1] = cvi(STRUCTURE)
pvmGLOB_old_time[1] = n<682246>
pvmGLOB_etime[1] = n<682246> / hz<100>
i++;
csi.number$ = i<2>
refresh$(csi(STRUCTURE))
pvmGLOB_old_cpu_sysinfo[2] = csi(STRUCTURE)
cvi.number$ = i<2>
refresh$(cvi(STRUCTURE))
pvmGLOB_old_cpu_vminfo[2] = cvi(STRUCTURE)
pvmGLOB_old_time[2] = n<682246>
pvmGLOB_etime[2] = n<682246> / hz<100>
i++;
csi.number$ = i<3>
refresh$(csi(STRUCTURE))
pvmGLOB_old_cpu_sysinfo[3] = csi(STRUCTURE)
cvi.number$ = i<3>
refresh$(cvi(STRUCTURE))pvmGLOB_etime[3] = n<682246> / hz<100>
i++;
old_vminfo = kstat$vminfo(STRUCTURE)
old_sysinfo = kstat$info(STRUCTURE)
initialized = <1>
last_number = number$<0>
return ;
ncpus = sysconf(<15>)
if (initialized<1> == <0>)
if (ncpus<4> > pvmGLOB_cpu_size<4>)
i = number$<0>
if (i<0> < <0> || i<0> >= ncpus<4>)
instance = pvmGLOB_old_cpu_sysinfo[0].instance$<0>
new_time = kstat$misc.clk_intr<682246>
itime = new_time<682246> - pvmGLOB_old_time[0]<682246>
if (itime<0> == <0>)
if (i<0> != last_number<0>)
return ;
i++;
p_vmstat$GLOBAL_pvm.number$ = i<1>
GLOBAL_pvm[1] = p_vmstat$GLOBAL_pvm(STRUCTURE)
ncpus = sysconf(<15>)
if (initialized<1> == <0>)
if (ncpus<4> > pvmGLOB_cpu_size<4>)
i = number$<1>
if (i<1> < <0> || i<1> >= ncpus<4>)
instance = pvmGLOB_old_cpu_sysinfo[1].instance$<1>
new_time = kstat$misc.clk_intr<682246>
itime = new_time<682246> - pvmGLOB_old_time[1]<682246>
if (itime<0> == <0>)
if (i<1> != last_number<0>)
smtx = pvmGLOB_cpu_sysinfo[1].mutex_adenters<0> / pvmGLOB_etime[1]<6822.46>
interrupts = pvmGLOB_cpu_sysinfo[1].intr<0> / pvmGLOB_etime[1]<6822.46>
system_calls = pvmGLOB_cpu_sysinfo[1].syscall<0> / pvmGLOB_etime[1]<6822.46>
context_switches = pvmGLOB_cpu_sysinfo[1].pswitch<0> / 
pvmGLOB_etime[1]<6822.46>invol_switches = 
pvmGLOB_cpu_sysinfo[1].inv_swtch<0> / pvmGLOB_etime[1]<6822.46>user_time 
= (<100> * pvmGLOB_cpu_sysinfo[1].cpu[1]<0> / (hz<100> * 
pvmGLOB_etime[1]<6822.46>))
system_time = (<100> * pvmGLOB_cpu_sysinfo[1].cpu[1]<0> / (hz<100> * 
pvmGLOB_etime[1]<6822.46>))
wait_time = (<100> * pvmGLOB_cpu_sysinfo[1].cpu[1]<0> / (hz<100> * 
pvmGLOB_etime[1]<6822.46>))
idle_time = (<100> * pvmGLOB_cpu_sysinfo[1].cpu[1]<0> / (hz<100> * 
pvmGLOB_etime[1]<6822.46>))
pages_in = pvmGLOB_cpu_vminfo[1].pgpgin<0> / pvmGLOB_etime[1]<6822.46>
pages_out = pvmGLOB_cpu_vminfo[1].pgpgout<0> / pvmGLOB_etime[1]<6822.46>
swapins = pvmGLOB_cpu_vminfo[1].swapin<0> / pvmGLOB_etime[1]<6822.46>
swapouts = pvmGLOB_cpu_vminfo[1].swapout<0> / pvmGLOB_etime[1]<6822.46>
scan = pvmGLOB_cpu_vminfo[1].scan<0> / pvmGLOB_etime[1]<6822.46>
pgrrun = pvmGLOB_cpu_vminfo[1].pgrrun<0> / pvmGLOB_etime[1]<6822.46>
last_number = i<1>
return ;
i++;
pvmGLOB_old_cpu_vminfo[3] = cvi(STRUCTURE)
pvmGLOB_old_time[3] = n<682246>



The only thing that is different on this machine than all the other 
machines that work is that it runs EMC BCV's (splitting, and syncing), 
which really means I have lots of SAN disk on this machine, some of 
which are offline (via veritas), and others are online.  Also running 
orca on this machine is a hudge CPU and memory hog.  If anyone has any 
ideas please let me know.

Jeff