Adding Journal Device to an existing node.
This section illustrates how to add a journal device to an existing node.
1. Enter maintenance mode
In order to add journal device to a node, the node must be put in maintenance mode:
pxctl service maintenance --enter
This is a disruptive operation, PX will restart in maintenance mode.
Are you sure you want to proceed ? (Y/N): Y
Check if the node is in maintenance mode:
PX is in maintenance mode. Use the service mode option to exit maintenance mode.
Node ID: a0b87836-f115-4aa2-adbb-c9d0eb597668
Local Storage Pool: 0 pool
Pool IO_Priority Size Used Status Zone Region
No storage pool
Local Storage Devices: 0 device
Device Path Media Type Size Last-Scan
No storage device
total - 0 B
Cluster ID: bb4bcf13-d394-11e6-afae-0242ac110002
Node IP: X.X.X.185 - Node ID: a0b87836-f115-4aa2-adbb-c9d0eb597668 In Maintenance
Global Storage Pool
Total Used : 0 B
Total Capacity : 0 B
AlertID Resource ResourceID Timestamp Severity AlertType Description
39 CLUSTER a56a4821-6f17-474d-b2c0-3e2b01cd0bc3 Jan 8 06:01:22 UTC 2017 ALARM Node state change Node a56a4821-6f17-474d-b2c0-3e2b01cd0bc3 has an Operational Status: Down
48 NODE a0b87836-f115-4aa2-adbb-c9d0eb597668 Jan 8 21:45:25 UTC 2017 ALARM Cluster manager failure Cluster Manager Failure: Entering Maintenance Mode because of Storage Maintenance Mode
2. Add the journal device to the node
A journal device is recommended to be an SSD/NVME:
pxctl service drive add --drive /dev/nvme01 --journal
Successfully added journal device /dev/nvme010p1
The journal device is expected to not have any filesystem on it, This can be verified by running
blkid /dev/nvme01. If it has a filesystem on it and you still want to use it as a journal device, the filesystem should be removed by running
wipefs -a /dev/nvme01.
3. Exit maintenance mode
pxctl service maintenance --exit