Acesso ao Cluster

Cluster – Comandos de Uso

Administração:

Admin – Config. Para desligar remoto
chmod a+s /sbin/shutdown
ethtool eth0
ethtool -s eth0 wol g
echo ‘/usr/sbin/ethtool -s eth0 wol g’ >> /etc/rc.d/rc.loca

Tutorial – Inclusão de novas máquinas

Comandos para ligar nodos remotamente

• Para ligar os nodos remotamente, basta logar no servidor labtmc e digitar no terminal o comando correspondente para:

  Ligar cpus quadcore (0 a 6)

  ligarpc0

  ligarpc1

  ligarpc2

  ligarpc3

  ligarpc4

  ligarpc5

  ligarpc6

• Para desligar os nodos remotamente, basta logar no respectivo nodo e digitar no terminal o comando desligar

Utilização do comando screen

Criar screen:

screen -S nomedascreen

Listar screen:

screen -ls

Abrir novas janelas dentro do screen:

crtl + a + c

Alternar as janelas abertas na screen:

crtl + a + 0

crtl + a + 1

…

crtl + a + 9

Desanexar screen

crtl +a + d

Desanexar screen (via linha de comando)

screen -d nomedascreen

Reanexar screen (via linha de comando)

screen -r nomedascreen

Modo scroll em uma screen

crtl + a+ esc (somente esc para sair)

 

Cluster Open MP

Expensão do uso da memória

• Para expandir o uso da memória, no terminal faça:

ulimit -s 2048000

export KMP_STACKSIZE=2048000000

(É necessário refazer o comando a cada login.)

• Para rodar programas que exigem muita memória RAM,  faça:

ifort -mcmodel=XXX  -shared-intel program.f

onde XXX pode ser  “small”, “medium” ou “large” . O comando  -shared-intel é usado para processadores intel.

 

Tutoriais:

• Tutorial MPI
• Tutorial Rodar com Bibliotecas MKL
• Tutorial Incluir novas máquinas

Como configurar o Cluster

Cluster install:

Modo AHCI (Trocar para ATA após instalado)

OBS: DVD interno não funciona. Usar Gaveta USB

OBS2: Se teclado usb não funciona durante install, usar tecl. Ps2

Pacotes Rocks:

NAME VERSION ARCH ENABLED

sge: 6.1.1 x86_64 yes ( job queueing system)

os: 6.1.1 x86_64 yes (required) CentOS 6.5 w/updates pre-applied

kernel: 6.1.1 x86_64 yes (required) Rocks Bootable Kernel

ganglia: 6.1.1 x86_64 yes Cluster monitoring system from UCB

web-server: 6.1.1 x86_64 yes (Rocks Web Server Roll)

area51: 6.1.1 x86_64 yes (System security related services and utilities)

base: 6.1.1 x86_64 yes (required) Rocks Base Roll

hpc: 6.1.1 x86_64 yes (Rocks HPC Roll)

Desabilitar Hard Boot:

How do I disable the feature that reinstalls compute nodes after a hard reboot?

When compute nodes experience a hard reboot (e.g., when the compute node is reset by pushing the power button or after a power failure), they will reformat the root file system and reinstall their base operating environment.

To disable this feature:

Login to the frontend

Create a file that will override the default:

# cd /export/rocks/install

# cp rocks-dist/arch/build/nodes/auto-kickstart.xml \

site-profiles/6.1.1/nodes/replace-auto-kickstart.xml

Where arch is “i386” or “x86_64”.

Edit the file site-profiles/6.1.1/nodes/replace-auto-kickstart.xml

Remove the line:

<package>rocks-boot-auto<package>

Rebuild the distribution:

# cd /export/rocks/install

# rocks create distro

Reinstall all your compute nodes

Note

An alternative to reinstalling all your compute nodes is to login to each compute node and execute:

# /etc/rc.d/init.d/rocks-grub stop

# /sbin/chkconfig –del rocks-grub

Instalar freenx server:

Currently there is a version of NX and FreeNX in the CentOS Extras repository for both CentOS 5 and CentOS 6.

nano /etc/yum.repos.d/CentOS-Base.repo

To install the stable version of NX / FreeNX, issue this command from the server:

[root@server ~]# yum install nx freenx

Admin – Config. Para desligar remoto:

chmod a+s /sbin/shutdown

ethtool eth0

ethtool -s eth0 wol g

echo ‘/usr/sbin/ethtool -s eth0 wol g’ >> /etc/rc.d/rc.local

Periódicos 

• arXiv
• NASA ADS
• IOPscience
• Web of Science
• Physical Review