The alert is triggered when a physical disk is inserted into or removed from a host.
Host { .labels._hostname }: The disk { .labels.disk_id_serial }({ .labels._device }) was unplugged at { .labels.created_time }.
Host { .labels._hostname }: The disk { .labels.disk_id_serial }({ .labels._device }) was unplugged at { .labels.created_time }. The disk was located on the slot { .labels._physical_slot } of brick { .labels._brick_name }.
Host { .labels._hostname }: The disk { .labels.disk_id_serial }({ .labels._device }) was plugged into host at { .labels.created_time }.
If the disk is an idle physical disk, there is no actual impact on system operations.
If the disk removed is a physical disk in use, this may cause:
The data replicas on the disk enter an abnormal state, triggering data recovery, which degrades cluster storage performance.
If the disk contains the operating system partition, the operational safety of the node is significantly reduced, and the node is at risk of failure.
If all physical disks containing the operating system are removed, the node will crash immediately.
If multiple nodes simultaneously experience the removal of physical disks in use before data recovery is complete, data may become inaccessible during the removal period and result in permanent data loss when the node or service is subsequently restarted after failure.
The operating system detects a disk insertion or removal and triggers the alert.
If this is a planned operation on an idle physical disk, manually mark the alert as resolved.
If a physical disk in use is accidentally removed, immediately re-insert it into the node. Because the system cannot automatically determine whether the disk went offline due to manual removal or a disk fault, it will not automatically mount the disk, to avoid repeatedly writing new data to a faulty area and causing greater risk. If you confirm that the disk itself is normal, run zbs-node set_disk_healthy /dev/xxx to mark the disk as healthy, then remount it through the AOC management interface.
If the physical disk was not actually removed but the alert still appears, collect kernel and IPMI logs and submit them to an Arcfra after-sales engineer or the hardware vendor for investigation. Common hardware fault causes include:
Physical disk failure
Disk backplane or power module anomaly
HBA card or RAID card anomaly
