Cache Drive Fault

The BoostIO distributed cache layer uses NVMe SSDs as tier-2 cache media to persist object data in the write and read caches. The methods for troubleshooting cache drives are described as follows.

**Table 1** Cache drive fault scenarios
Scenario	Impact	Handling Method	Remarks
Cache drive fault	The write cache does not allow data writes. Object data of the write cache stored on drives is lost. Read cache data stored on drives cannot be read.	Detects drive faults and reports the faults to the cluster management module for faulty drive isolation. Partition data affected by drive faults is forcibly evicted to the back-end storage.	Service I/Os are retried (maximum retry time: 60 seconds) during the period from drive fault occurrence to partition view switchover. In this case, foreground I/Os are interrupted. This version does not support slow/faulty drive recovery, dirty drive reinsertion, hot plug, and single-drive expansion.

Parent topic: Troubleshooting