IBM GC28-1982-02 manual 0031-307 Ÿ, Initialization, Reason, Restart

0031-307 Ÿ 0031-311

0031-307	remote	child:	error restoring stdin.
Explanation:		The previously closed stdin cannot be restored.
User Response:		Probable system error. Gather information about the problem and
local	site	procedures		for reporting hardware and software problems.
0031-308	Invalid	value	for	string: string

Explanation: Indicated value is not a valid setting for the indicated environm command line option.

User Response: Set to a valid value and rerun.

0031-309

Connect

failed

during

message

passing

initialization,

task

number,

reason:

string

Explanation:

The

Communication Subsystem was unable to

connect this task to one

o

more

other

tasks

in

the

current

partition

for

the

reason

given.

User

Response:

If

a

timeout

has

occurred,MP TIMEOUT the environment

variable

is

set

to too

low

of

a value. (The default value is 150 seconds.) If you have n

MP_TIMEOUT

environment

variable

and

the

program

being

run

under

POE

is

NFS

mount

150

seconds

may

not

be

sufficient.

If

the

reason

given indicates "Permission denied", you

should

ensure

the

log

user ID of the user

submitting the job is consistent

on

all

nodes

on

which

If

the

reason

given

indicates

"Permission

denied"

or "Not owner" and the job

under LoadLeveler, you should ensure that the adapter requirement given to L

compatible with MP_theEUILIB

environment

variable.

If

the

reason

given

indicates

"No

such

device",

the

Communication

Subsystem

li

(libmpci.a

)

bound

into the executable does not match the

switch

adapter

for

error usually occurs when the executable was statically bound on a system

configured

for

a different switch adapter. For

example, a program that was

system

configured

with

a TB2 adapter, and was then

attempted

to

be

run

on

TB3

adapter.

In

this

case,

you

should

recompile

the program on a system c

same

switch

adapter

as

that

of

node

where

the

executable

will

be

run.

For

any

other

reason,

an internal error has occurred.

You

should

gather

info

the

problem

and

follow

local

site

procedures

for

reporting

hardware

and

sof

0031-310

Socket

open

failed

during

message

passing

initialization,

task

number,

reason:

string

Explanation:

The

Communication

Subsystem

was

unable

to

open

a

socket

for

message

passing

for

the

indicated

task

for

the

reason

given.

User

Response:

If the reason given is “No buffer space available,” have the

administrator raise

thesb valuemax

ofusing

theno

command.

The

current

suggested

value is

128000.

For

any

other

reason,

an internal error has most

likely

occurred.

Gather

info

the

problem

and

follow

local

site

procedures

for

reporting

hardware

and

sof

0031-311

Restart

of

program

string failed.

Return code

is

number.

Explanation:

The

restart of the program indicated

was

unsuccessful.

User

Response:

Check

that

the

program

name

is

valid,

and

that

it

was

previousl

checkpointed.

Chapter 4. POE Messages71