A DM-multipath device has remained suspended for 1 minute. The node-exporter
dmmultipath collector reports node_dmmultipath_device_active as 0 for a
suspended device and 1 for an active device.
I/O through the device is blocked. Applications and VMs using it may hang or experience I/O timeouts. This alert does not establish why the device was suspended.
Use the alert’s instance, device, and sysfs_name labels to identify the
node and device. Confirm the node corresponding to the node-exporter scrape
target. On that node, inspect the device and multipath status:
dmsetup info <device>
dmsetup status <device>
multipath -ll <device>
journalctl -u multipathd
Check whether a storage management operation intentionally suspended the device, and inspect underlying paths, storage connectivity, and system logs. Coordinate with the administrator or component that manages the device.
Check the device state in Prometheus:
node_dmmultipath_device_active{job="node-exporter",instance="<instance>",device="<device>",sysfs_name="<sysfs_name>"}
Resolve the underlying storage or management issue before resuming the device. Coordinate restoration with the device’s managing component or administrator; a suspended device can be part of an in-progress storage operation.
Verify that the device is active, workload I/O resumes, and
node_dmmultipath_device_active returns to 1. Do not remove or recreate the
device while workloads are using it.
If you cannot resolve the issue, see the following resources: