Going through the code in trunk, I think your observation is right. On Mon, Jul 8, 2013 at 11:22 AM, Thanh Do <[email protected]> wrote:
> hi all, > > As far as I understand, there are two ways > that a memstore flush is triggered. > The first one is via MemstoreFlusher thread, > and this is a background work for sure. > > Second, a client can trigger a flush by using > flushRegion API. It seems to me that this API > is executed synchronously, meaning the memstore > flush in this case is essentially a foreground workload, > in the sense that it requires one IPC handler to execute > the request and the handler only returns when > the flush is complete. > > Am I correct here? > > Thanks for your clarification. > > Thanh >
