SUMMARY: PBS (Portable Batch System)

From: Padiyath Kumar <Kumar.Padiyath_at_psi.ch>
Date: Sat, 05 May 2001 18:03:31 +0200

     I could solve the problem with a tip from Ralf Eichmann.Thanks to
Thomas Blinn
     and Ralf Eichmann.
     I was defining one parameter path wrongly with configure command.
     I donot have a file system /var. var is a directory in /usr file
system.
     The solution is:

     configure --set-server-home=/usr/var/spool/.. .
     instaed of
     configure --set-server-home=/var/spool/..

     The directory /usr/var/spool/.. is used to store all the necessary
     parameters to start the daemon.

  The reply from Ralf Eichmann:
  -----------------------------

what is your PBS_HOME directory? Is it somewhere inside /var? Note, that
on most Tru64 systems, var ist linked to usr/var in /. This leads to
some endless recursion during startup of pbs_server, pbs_mom and
pbs_sched, when these try to find the working directory. I forgot the
name of the function, I did the debugging some time ago.
The solution for us was to specify --set-server-home=/usr/var/...
instead of --set-server-home=/var/... for ./configure.

HTH,
Ralf Eichmann

The reply from Thomas Blinn:
========================

Your max-per-proc-stack-size is larger than your per-proc-stack-size,
so if you unlimit the process (e.g., ulimit command or whatever, it
depends on your shell), you will get a somewhat larger stack size.

But that will only help if it's really perfectly normal for this app
to need more than a 2048 kbytes stack size. If the application has a
bug that makes it recurse forever, for example, then increasing the
stack size is likely to only make it run a while longer before it just
dies again.

You need to get to the root cause of why the app consumed a 2048K stack
and still wanted more.

Did this app EVER work before on this or any other system?

 

My Question:
============

      I get the following error when I start pbs_server on a Tru64 work
station with the following specifications:
      If it is a stack problem how can I increase it?


     Error:
     ====
    # pbs_server -t create

      "Stack overflow: pid 2790, proc pbs_server, addr 0x11dffff30, pc
0x12003909c
      Segmentation fault (core dumped)"


      The configuration of the machine:
      ========================
   limit command output:
    ================

cputime unlimited
filesize unlimited
datasize 131072 kbytes
stacksize 2048 kbytes
coredumpsize unlimited
memoryuse 86376 kbytes
vmemoryuse 1048576 kbytes
descriptors 4096

    sysconfig -q proc output:
    ==================

# sysconfig -q proc
proc:
max-proc-per-user = 64
max-threads-per-user = 256
per-proc-stack-size = 2097152
max-per-proc-stack-size = 33554432
per-proc-data-size = 134217728
max-per-proc-data-size = 1073741824
max-per-proc-address-space = 1073741824
per-proc-address-space = 1073741824
autonice = 0
autonice-time = 600
autonice-penalty = 4
open-max-soft = 4096
open-max-hard = 4096
ncallout_alloc_size = 8192
round-robin-switch-rate = 0
round_robin_switch_rate = 0
sched-min-idle = 0
sched_min_idle = 0
give-boost = 1
give_boost = 1
maxusers = 32
task-max = 277
thread-max = 552
num-wait-queues = 64
num-timeout-hash-queues = 32
enable_extended_uids = 0
enhanced-core-name = 0
enhanced-core-max-versions = 16

    Operating System:
    ==============

    Tru64 V4.0D

   swapon -s output:
   =============

   # swapon -s
Swap partition /dev/rz3b (default swap):
    Allocated space: 15872 pages (124MB)
    In-use space: 441 pages ( 2%)
    Free space: 15431 pages ( 97%)

Swap partition /dev/rz2b:
    Allocated space: 15872 pages (124MB)
    In-use space: 488 pages ( 3%)
    Free space: 15384 pages ( 96%)


Total swap allocation:
    Allocated space: 31744 pages (248MB)
    Reserved space: 3930 pages ( 12%)
    In-use space: 929 pages ( 2%)
    Available space: 27814 pages ( 87%)

Total Memory: 96Mbytes


     with regards,
     Kumar
Received on Sat May 05 2001 - 16:04:52 NZST

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:42 NZDT