If we fail to allocate storage, it can be because the request is
too large to serve in one bite, so chop it down until we reach fetch_chunksize before we start LRU flushing to make space for it. Fixes #880
Showing
Please register or sign in to comment