On 05/01/2013 03:42 AM, Amos Jeffries wrote: > On 1/05/2013 10:21 a.m., babajaga wrote: >> Am I correct in the following: >> The selection of one of the 2 cache_dirs is not deterministic for same URL >> at different times, both for round-robin or least load. The selection of a cache_dir to _store_ a new object may indeed be nondeterministic in this sense. >> Which might have the consequence of generating a MISS, although the >> object ist cached in the other cache_dir. The selection of a disk to _load_ a cached object is always deterministic. The cached object will be found and loaded from the correct cache_dir. > Your whole chain of logic above depends on the storage areas (cache_dir) > being separate entities. This is a false assumption. They are only > separate to the operating system. They are merged into a collective > "cache" index model in Squid memory - a single lookup to this unified > store indexing system finds the object no matter where it is (disk or > local memory) with the same HIT/MISS result based on whether it exists > *anywhere* in at least one of the storage areas. Amos description is correct on a high level, but, just to avoid any future wrong assumptions, modern Squid may use separate indexes for some cache_dirs and do separate lookups in those cache_dir indexes. However, the resulting lookup speed is probably almost the same and a "cached object will be found at the right disk" principle still applies, which is all that matters for this high-level discussion. HTH, Alex.
