fifi cluster manual
[root@headnode~]# vim /etc/pam.d/system-auth
...
6 auth required /lib64/security/pam_tally2.so deny=5 unlock_time=10 no_magic_root
...
13 account required /lib64/security/pam_tally2.so no_magic_root reset
...
19 password requisite pam_pwquality.so try_first_pass local_users_only retry=8 authtok_type= minlen=8 lcredit=-1 ucredit=-1 dcredit=-1 ocredit=-1
...
[root@headnode~]# vim /etc/login_defs
#PASS_MAX_DAYS=90
#PASS_MIN_DAYS=1
[root@headnode~]# visudo
root ALL=(ALL) ALL
<SUDOERS_ID1> ALL=(ALL) ALL
<SUDOERS_ID2> ALL=(ALL) ALL
[root@headnode~]# vim /etc/group
wheel:x:10:root, <SUDOERS_ID1>, <SUDOERS_ID2>
[root@headnode~]# vim /etc/pam.d/su
uncomment following lines
auth required pam_wheel.so use_uid
[root@headnode~]# chgrp wheel /bin/su
[root@headnode~]# chmod 4750 /bin/su
[root@headnode~]# yum install -y squid
[root@headnode~]# systemctl start, enable squid
[root@headnode~]# vim /etc/squid/squid.conf
acl localnet src 10.10.10.0/24 # nodes
acl localnet src 192.168.10.0/24 # nodes-nfs
acl SSL_ports port 443
acl Safe_ports port 80 # http
# acl Safe_ports port 21 # ftp
acl Safe_ports port 443 # https
...
http_access deny !Safe_ports
http_access allow localhost manager
http_access deny manager
...
http_access allow localnet
http_access allow localhost
...
http_access deny all
[root@headnode~]# systemctl restart squid
[root@node07~]# /etc/yum.conf
proxy=http://10.10.10.3128
[root@node13~]# umount /home, /appl, /data
[root@node13~]# mount -a
or
[root@node13~]# umount /home, /appl, /data
[root@node13~]# mount -t nfs headnode:/home /home
[root@node13~]# mount -t nfs headnode:/appl /appl
[root@node13~]# mount -t nfs headnode:/data /data
[root@node13~]# mount -t nfs cs01:/project /project
You need to 1. create the linux account 2. update NIS 3. update SLURM_DB
[root@headnode~]# adduser <USER ID>
[root@headnode~]# passwd <USER ID>
...
...
[root@headnode~]# make -C /var/yp
[root@headnode~]# su <USER ID>
[USER_ID@headnode~]$ ssh-keygen -t rsa
...
[USER_ID@headnode~]$ cd ~/.ssh
[USER_ID@headnode~]$ cat id_rsa.put >> authorized_keys
Plug out the infiniband of snoopy
headnode. Reboot the node21~node40
. Replug the snoopy
infinband cable.
To do same execution throuout entire calculation nodes, do following.
[root@headnode~]# for i in {1..40}; do ssh node$(printf "%02d" $i) echo < hostname; done
[root@headnode~]# for i in {1..10}; do ssh c$(printf "%02d" $i) echo < hostname; done
cat /proc/cpuinfo
To check the current cpu temperature (to check the the air cooling in the server room without going)
[root@headnode~]# yum install lm_sensors
[root@headnode~]# for i in {1..40}; do ssh node$(printf "%02d" $i) yum install -y lm_sensors
...
[root@headnode~]# sensors
[root@headnode~]# scontrol update partitions=G1 state=Drain reason=Maintenance
[root@headnode~]# cat /var/lib/slurm/slurmctld.log | grep error
...
[root@node01~]# cat /var/lib/slurm/slurmd.log | grep error
CHECK USER LIST
[root@headnode~]# sacctmgr list user -s
ADD
[root@headnode~]# sacctmgr add user <USER_ID> account=phyiscslab
Example
[root@headnode~]# sacctmgr list qos
[root@headnode~]# sacctmgr modify qos normal set maxtresperuser=cpu=120