Big Data

Problem with SSD corruption on power failure .. flactuation

only enterprise-class SSDs can be relied upon (at all) for safe behavior on power fail.

The enterprise-class SSDs have super-capacitors that store enough power to write all data stored in the RAM within the SSD on power fail, and vendors charge three times as much as they do for consumer class SSDs.

Some vendors do certify their SSDs, but you should check them under real power-fail conditions, like pulling the plug.

It is better to use an Arduino with a relay for experimenting with cutting off power.


Node cannot be added to the GPFS cluster


There is an indication leading you to the conclusion that a node cannot be added to a cluster and steps to follow to correct the problem.

That indication is:

  • You issue the mmcrcluster or mmaddnode command and receive the message:
    Node nodeName was not added to the cluster. The node appears to already belong to a GPFS™ cluster.
Steps to follow if a node cannot be added to a cluster:

  1. Run the mmlscluster command to verify that the node is not in the cluster.
  2. If the node is not in the cluster, issue this command on the node that could not be added:
    mmdelnode -f
  3. Reissue the mmaddnode command.


Quickly mount a NTFS disk read/write on all recent OS X (including Yosemite, El Capitan):

  1. Open Application -> Utilities -> Terminal
  2. Type mount and look for the line with your disk. It will show something like: /dev/disk3s1 on /Volumes/MyDisk (ntfs, local, noowners, nobrowse)
  3. Type the following in the Terminal, replacing /dev/diskXsX with your disk as shown in the mount command:
    sudo mkdir /Volumes/Mount  
    sudo mount -o rw,auto,nobrowse -t ntfs /dev/diskXsX /Volumes/Mount/
    open /Volumes/Mount/

To make this change permanent run the following (correct for El Capitan):

  1. Run the following command, changing /dev/diskXsX to your disk:
    export DEVICE=/dev/diskXsX   
    echo UUID=`diskutil info $DEVICE | grep UUID | awk '{print $3}'` none ntfs rw,auto,nobrowse 
    echo LABEL=NTFS none ntfs rw,auto,nobrowse
  2. Run sudo vifs and paste in the output from the previous 2 lines. To do this press down to go to the bottom of the file, A to start adding text, paste in the 2 lines then press escapeand :wq to write the file. (vifs is the only safe way to edit the fstab in OS X).
  3. Run the following to mount the disk
    sudo umount /Volumes/$DEVICE 
    sudo diskutil mountDisk $DEVICE
    sudo open `mount | grep $DEVICE | awk '{print $3}'`

Note: The device will no longer automatically open a window when you attach it. To access it open a Finder window and select the Go -> Go to Folder to /Volumes


Creating JBOD with Areca controller, for Ceph

How to create JBOD with Areca controller, it is very easy .

Since redundancy and failure maintain by ceph itself , it is better to have disks with JBOD instead of Raid when integrated with CEPH.

To use Areca in JBOD:
– enter RAID setup by pressing Tab/F3 when the Areca BIOS bootstrap tells you so
– remove all RAID sets including any passthrough disks
– go to “raid functions” menu and choose “JBOD/RAID mode” and set it to JBOD instead.
– go back to main menu and press F10 to reboot


Yea, or you can do it through the web management console.
System Controls -> system configuration.



How does ISCSI works ?

ISCSI involves two main factors, 1st the ISCSI initator (client) and 2nd ISCSI target (server) which is most often a disk server running any avaiable ISCSI target daemon . ISCSI can be used to mount network drives locally and use as normal disk drives in Linux sense they can be added as /dev/sdX for e.g

Various cloud platform such as OpenStack use it to provide Elastic block storage using . The block storage service of Openstack named as Cinder. (Release Havana, Grizzly).

Further Reading:




Using Google Spreadsheets as a Database in the Cloud