Skip to content

Latest commit

 

History

History

simplemultilinux-swapdisks

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 
 
 
 
 

Multi Node Linux Cluster

This demonstrates booting multiple linux machines and shows how to setup an experiment to cause the OS drive to land on SDB1.

Repro to swap disks

  1. deploy the templateFile using your CLI or clicking deploy to portal above. Choose one of the multi core machines

  2. SSH to the first machine on port 2200 (second machine is on port 2201, 3rd on port 2202, and so on)

  3. You need to place your private key in /home/azureuser/.ssh

  4. mkdir .ssh

  5. cd .ssh

  6. vi id_rsa to edit your file and paste in the contents of your private key

  7. chmod 600 ~/.ssh/id_rsa

  8. paste in the contents of scandrives.sh and findr.sh into /home/azureuser

  9. chmod +x /home/azureuser/scandrives.sh

  10. chmod +x /home/azureuser/findr.sh

  11. findr.sh

  12. wait until the script breaks out with a repro.

Repro to show /etc/fstab gets written incorrectly

  1. deploy the templateFile using your CLI or clicking deploy to portal above. Choose one of the multi core machines

  2. SSH to the first machine on port 2200 (second machine is on port 2201, 3rd on port 2202, and so on)

  3. You need to place your private key in /home/azureuser/.ssh

  4. mkdir .ssh

  5. cd .ssh

  6. vi id_rsa to edit your file and paste in the contents of your private key

  7. chmod 600 ~/.ssh/id_rsa

  8. paste in the contents of scandrives.sh into /home/azureuser

  9. chmod +x /home/azureuser/scandrives.sh

  10. ./scandrives.sh --get-fstab to see the values in /etc/fstab of all machines, and observe /dev/disk/azure/resource-part1 refers to /mount

  11. (steps 1-7 must be done within first 10 minutes after cluster has deployed) './scandrives.sh --reboot-nodes' to reboot all machines

  12. after 30 seconds ./scandrives.sh --get-fstab and notice how all /etc/fstab files have been re-written and /dev/sdb1 refers to /mount

Repro to get NTFS mounted ephemeral disk

  1. deploy the templateFile using your CLI or clicking deploy to portal above. Choose at least 20 machines with 16.04.0-LTS.

  2. SSH to the first machine on port 2200 (second machine is on port 2201, 3rd on port 2202, and so on)

  3. You need to place your private key in /home/azureuser/.ssh

  4. mkdir .ssh

  5. cd .ssh

  6. vi id_rsa to edit your file and paste in the contents of your private key

  7. chmod 600 ~/.ssh/id_rsa

  8. paste in the contents of scandrives.sh and findr.sh into /home/azureuser

  9. chmod +x /home/azureuser/scandrives.sh

  10. ./scandrives.sh --get-mounts | grep fuse - this may show an NTFS mounted drive

  11. Edit the template to choose a different VM Size, and re-deploy

  12. once deployed, run ./scandrives.sh --get-mounts | grep fuse to show the NTFS mounted drives