If compression/decompression fails during block reclaiming, corrupted data will be stored internally, and then it causes crash in the next round block reuse or file re-open.
We should handle any kinds of errors around block reclaiming. If error happens, it should be propagated to upper layers so as to avoid the situation that block reuse logic is running with corrupted data.
|For Gerrit Dashboard: MB-23583|
|76920,3||MB-23583: Gracefully handle errors during block reclaiming||spock||forestdb||Status: MERGED||+2||+1|
|77072,2||[BP] MB-23583: Gracefully handle errors during block reclaiming||stable||forestdb||Status: MERGED||+2||+1|
|80332,2||[BP] MB-23583: Gracefully handle errors during block reclaiming||watson||forestdb||Status: MERGED||+2||+1|