Saturday, February 21, 2015

ubuntu MAAS


1. configure the IP address of MAAS server: (the second one works)
sudo dpkg-reconfigure maas-cluster-controller
sudo dpkg-reconfigure maas-region-controller    

weird DELL's IPMI (or IDRAC, whatever)

the –o intelplus is an additional command line option / OEM specifier. Please try it together with the other options, e.g.

ipmitool –I lanplus –U user –P password –H host –o inteloem mc info

should give you the same information as with an IPMI 1.5 connection.

ipmitool –I lan –U user –P password –H host –o inteloem mc info

If this works you have RMCP+ working and you can try to activate sol



ipmitool –I lanplus –U user –P password –H host –o inteloem sol activate


I'm able to use ipmitool with a remote machine using the 'lan' protocol, but 'lanplus' fails for every command. I'm trying to get SOL working, but I'm having no luck with lanplus at all:

ipmitool -v -I lanplus -E -H 10.0.40.240 -U admin lan print

Error in open session response message : insufficient resources for session

Error: Unable to establish IPMI v2 / RMCP+ session

I can ping the remote BMC with both ipmiping and rmpcping

The machine is a Dell CS23, the BMC has IPMI v2. I'm using ipmitool v1.8.12 compiled from source.
I've tried cold resetting the BMC, but it doesn't help.

Saturday, February 14, 2015

migrate KVM virtual machines

  1. Create the logical volume on the target. It should be the same size as the source volume.
  2. Run nc -l 1234 | dd of=/dev/<vgname>/<lvname> on the target host. This tells netcat to listen on the network port and pipe everything to the logical volume.
  3. Shutdown the source machine cleanly to write everything to disk.
  4. Run dd if=/dev/<vgname>/<lvname> | nc <targetip> 1234 on the source host. This tells netcat to pipe everything over the network to the target host.
  5. Go fetch a cup of coffee. This takes a few minutes depending on your volume size and network speed.
  6. Copy the xml definitions from /etc/libvirt/qemu/ from the source to the target.
  7. Edit the storage portion of the xml file and any other necessary bits such as network bridge name if they are different.
  8. Reload libvirt on the target to read the xml. An /etc/init.d/libvirt-bin reload should do the trick (or systemctl restart libvirtd for distributions that use systemd)
  9. Start-up the machine on the target host.

Saturday, January 10, 2015

build hadoop on traditional HPC

several useful tutorials:
1. http://www.glennklockwood.com/di/hadoop-deploy.php

Configuration:

DFS_NAME_DIR
DFS_DATA_DIR
MAPRED_LOCAL_DIR

config_subs[MASTER_NODE]="$MASTER_NODE"
config_subs[MAPRED_LOCAL_DIR]="$MH_SCRATCH_DIR/mapred_scratch"
config_subs[HADOOP_TMP_DIR]="$MH_SCRATCH_DIR/tmp"
config_subs[DFS_NAME_DIR]="$MH_SCRATCH_DIR/namenode_data"
config_subs[DFS_DATA_DIR]="$MH_SCRATCH_DIR/hdfs_data"
config_subs[DFS_REPLICATION]="$MH_DFS_REPLICATION"
config_subs[DFS_BLOCK_SIZE]="$MH_DFS_BLOCK_SIZE"
config_subs[MAPRED_TASKTRACKER_MAP_TASKS_MAXIMUM]="$MH_MAP_TASKS_MAXIMUM"
config_subs[MAPRED_TASKTRACKER_REDUCE_TASKS_MAXIMUM]="$MH_REDUCE_TASKS_MAXIMUM"
config_subs[MAPRED_MAP_TASKS]="$MH_MAP_TASKS"
config_subs[MAPRED_REDUCE_TASKS]="$MH_REDUCE_TASKS"
config_subs[HADOOP_LOG_DIR]="$MH_SCRATCH_DIR/logs"
config_subs[HADOOP_PID_DIR]="$MH_SCRATCH_DIR/pids"