replace the write disk

[wimpers]

REPLACING A WRITE DISK:

In this example we have a cluster of 3 ovs nodes with a active vpool running on them (called myvpool):

10.100.199.151 - ovs-node01
10.100.199.152 - ovs-node02
10.100.199.153 - ovs-node03

On node 10.100.199.151 - ovs-node01 we have a disk that is broken, to be more specific the WRITE disk on mountpoint /mnt/hdd3
As a result the volumedriver has stopped serving vdisks but the service can still be running.

PART 1: Stop the volumedriver and remove the broken disk

• Stop the volumedriver its service: stop ovs-volumedriver_<vpoolname>
• Remove the broken disk, and replace it

PART 2: Adding the new disk

• Detect the new disk in the system: df -h or lsblk
• Format, partition and put a filesystem on the disk
• Detect the newly created filesystem in directory: ls /dev/disk/by-id/ -l
• Edit /etc/fstab so /mnt/hdd3 is mapped on the new disk
• Perform command mount -a
• Check if the directory is mounted: df -h | grep hdd3
• Create directories on /mnt/hdd3:
  • _write_dtl_1
  • _write_fcache_1
  • _write_fd_1
  • _write_sco_1

PART 3: FINISHING UP

• Resync the GUI its disk roles on the storage router page
• Start the volumedriver services:
  • ovs-dtl_
  • ovs-volumedriver_
• Check if some services are down via: ovs monitor services
• Run the healthcheck