You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The text was updated successfully, but these errors were encountered:
Basho-JIRA
changed the title
Riak KV (w/LevelDB backend) running as a single node does not persist data
Riak KV (w/LevelDB backend) running as a single node does not persist data [JIRA: RIAK-2416]
Feb 27, 2016
Problem is within DB::Open(). This routine creates an internal DBImpl object, but continues to use original parameters past to the call. The DBImpl object contains the database name (database path) as corrected for tiered storage. DB:Open()'s use of the original database path is incorrect.
The problem is limited to the first recovery log, 0000xxxx.log, created upon database start/restart. All subsequent recovery logs get created using the proper tiered storage corrected path.
Result is that if database goes up and down before the initial recovery log is turned into a .sst table file, that log's data is lost. All subsequent recovery logs work properly during start/stop.
Fix is to adjust DB::Open() to use impl->dbname_ instead of dbname (parameter passed). Would be smart to switch all file operations after creation of DBImpl object to use impl->options_ instead of passed options parameter.
RiakKV in a single node, non-clustered setup does not seem to persist data to disk using LevelDB as a backend.
Overview of test:
Attached is the procedure I followed verbatim and was able to reproduce.
I do have the machine the test was performed on publicly available on EC2 if anyone wants to get in to inspect.
ProcedureToReproduce.txt
The text was updated successfully, but these errors were encountered: