On 27 Jan 2011, at 16:14, Andrew Deason wrote: >> Is there a way to measure how effective the chunksize is on a live >> system? > > You can look at the cache performance metrics, and see if your miss > ratios seem worse than before you changed it.
Chunksize does affect more than cache performance though. In particular, larger chunksizes will slow down non-chunk aligned file access. When we fetch files from the fileserver, we fetch by chunk - so if you are reading data at the end of a chunk, then the whole chunk will be fetched before your read completes. If you have a large chunksize, and a slow link, then this will impact performance. On Unix, the same is true when writing back to the server. Even if you only change a byte of data, we'll write back the whole chunk containing that byte. Fixing this so that the cache manager maintains information about the range of bytes that have been dirtied would be a relatively straightforward performance optimisation. Cheers, Simon. _______________________________________________ OpenAFS-info mailing list OpenAFS-info@openafs.org https://lists.openafs.org/mailman/listinfo/openafs-info
