hesela.devNVMe LBA Status Information log
The NVMe LBA Status Information log identifies namespace ranges that warrant further examination for potentially unrecoverable logical blocks.
Log 0Eh has controller scope, with elements for attached namespaces. Its range descriptors and recommended action (RATYPE) guide subsequent Get LBA Status commands; they are not those commands' returned status descriptors. [1]
Reading with RAE=1 freezes this log until a read with RAE=0 clears the outstanding alert, restarts its report interval and permits updates. This snapshot rule is specific to this log, not a universal property of RAE. [1]
For RATYPE=10h, acknowledgement suppresses the same namespace element unless another component failure may have created additional untracked LBAs. Retain the guidance before acknowledgement; clearing an alert does not repair data. [1]
Support is optional; check OACS.GLSS for the error-oriented capability rather than inferring support from allocation reporting alone. [1]
Compare the generation counter in the RAE=0 acknowledgement header with the original capture. A mismatch indicates newer log data and permits rereading without waiting for the usual poll interval; it is not evidence of completed recovery. [1]
Relations
Limitations
- NLRD=FFFFFFFFh means location information is unavailable and the whole namespace needs examination, not that no blocks are affected. [1]
- ESTULB=FFFFFFFFh means an unavailable estimate, not a measured count of potentially unrecoverable blocks. [1]
Citations
Sources & caveats
- Source interpretation of NVM Command Set 1.2; no controller experiment.