[mephi-hpc] Часть задач снимается при запуске на кластере Cherenkov

Alexey Bogdanov AGBogdanov at mephi.ru
Fri Jun 1 17:11:47 MSK 2018


Здравствуйте!

Возникла проблема - часть задач (практически идентичных) при запуске на 
кластере Cherenkov снимается, другая часть - нормально работает. Вот 
пример сообщения, приходящего при этом на e-mail.

 From adm at master.cherenkov  Fri Jun  1 16:30:09 2018
Return-Path: <adm at master.cherenkov>
X-Original-To: hydra at master
Delivered-To: hydra at master
Received: by master (Postfix, from userid 0)
     id 42799143C07; Fri,  1 Jun 2018 16:30:09 +0300 (MSK)
To: hydra at master
Subject: PBS JOB 54634.master
Precedence: bulk
Message-Id: <20180601133009.42799143C07 at master>
Date: Fri,  1 Jun 2018 16:30:09 +0300 (MSK)
From: adm at master.cherenkov (root)

PBS Job Id: 54634.master
Job Name:   EASNVD165508_x25a_5
Exec host:  n219/24
Aborted by PBS Server
Job cannot be executed
See Administrator for help

 From adm at master.cherenkov  Fri Jun  1 16:30:17 2018
Return-Path: <adm at master.cherenkov>
X-Original-To: hydra at master
Delivered-To: hydra at master
Received: by master (Postfix, from userid 0)
     id 4478B143C05; Fri,  1 Jun 2018 16:30:17 +0300 (MSK)
To: hydra at master
Subject: PBS JOB 54634.master
Precedence: bulk
Message-Id: <20180601133017.4478B143C05 at master>
Date: Fri,  1 Jun 2018 16:30:17 +0300 (MSK)
From: adm at master.cherenkov (root)

PBS Job Id: 54634.master
Job Name:   EASNVD165508_x25a_5
Exec host:  n219/24
An error has occurred processing your job, see below.
request to copy stageout files failed on node 'n219/24' for job 54634.master

Unable to copy file /var/spool/pbs/spool/54634.master.OU to 
/home/hydra/geant4.10.3/examples/basic/NVD91_Cascade-build/EASNVD165508_x25a_5.o54634
*** error from copy
*** end error output

Unable to copy file /var/spool/pbs/spool/54634.master.ER to 
/home/hydra/geant4.10.3/examples/basic/NVD91_Cascade-build/EASNVD165508_x25a_5.e54634
*** error from copy
*** end error output


Похоже, что один из узлов отклоняет задачи. Посмотрите, пожалуйста, в 
чем проблема!

Всего хорошего,
с уважением,
Алексей Богданов



More information about the hpc mailing list