On 27/10/2015 10:54 a.m., HackXBack wrote: > Facing the same problem, > by default if i didnt use range_offset_limit , idm download the file with > multiple mirros, all are 206 but cant be cached and hit when repeat the same > url download. > when i use range_offset_limit, idm download the file with 1 mirror, this > will decrease the speed but it can be hit when repeat the same url download. > > so users have problems in this , they use idm because they need to download > the file with multiple mirrors, and in the same time they need it to be hit > when repeat the download file. > > there must be a way to cache idm downloads and make the download with > multiple mirrors, right haa ? Not until someone patches Squid to do Range response caching. Meanwhile I question this "need". D/L from multiple mirrors and caching the result as one object breaks future D/L from having the same "need" met. Since the future responses are served from cache instead of those multiple mirrors. If it is acceptible for one fetch to be single-soruced. Then why is there a need for an identical other requets to be multiple servers. It does not make sense for the "need" to exist. Seems to me the "need" is actually a "want" and a greedy one at that. Remember that each browser/client fetching tens of requests. Means a proxy deals with requests by the thousand. Means an origin is possibly serving close to millions of endpoints with each response. Now multiple those numbers by 5... It is a better and more responsive system that conserves those upstream/server resources and sends as few requests up as necessary. Amos _______________________________________________ squid-users mailing list squid-users@xxxxxxxxxxxxxxxxxxxxx http://lists.squid-cache.org/listinfo/squid-users
