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
in db test changing dynamically write buffer size and it is not aligned to the fact we have an active memtable and a pending memtable...
so the check of the L0 files (meainng the memtable that have done flush) depending the increase/decrease write buffer size is incorrect cause the pending memtable is created with the previous write buffer size ...
The text was updated successfully, but these errors were encountered:
Found a way to fix it. i tested and it was solved.
i will transfer for review
the fix is in the code.... not in the test.
when i create the pending memtable i dont activate it with allocator resources that cause the issue (part of the flush threshold but it shouldn't , cause it doent in the game yet) when it turns active i set the current cf values to the activate memtable so if the write buffer size was changed after the pending memtable was created , it will get the correct value as part of activation and the test will get what it want.
note that the memtable object creation is not part of the flush threshold - the bucket shouldnt be calculate as a write buffer. so it doesnt use the arena. (it is like that today - just mention it to explain that its ok to set the arena parameters after the memtable was created but wasnt activated yet)
in db test changing dynamically write buffer size and it is not aligned to the fact we have an active memtable and a pending memtable...
so the check of the L0 files (meainng the memtable that have done flush) depending the increase/decrease write buffer size is incorrect cause the pending memtable is created with the previous write buffer size ...
The text was updated successfully, but these errors were encountered: