G'day, I've recently committed some changes to the delay pool code in squid-2.HEAD.
* Statistics for traffic transferred per-bucket and per-aggregate in each pool can be fetched via "cachemgr mgr:delay2". Its in a machine-friendly format (a la "cachemgr mgr:5min") and its useful primarily to make certain that your delay pool ACLs are placing traffic in the correct pools and buckets. * I've committed some work to include a "delayed" ACL format for delay pools. The client coded up a basic patch to allow replies to come in at full speed until a certain size was reached, then the server-side connection was placed in the delay pool. This way you can allow full-speed downloads for small objects and throttle large downloads after a certain size. The (initial!) changeset for this which describes what is going on can be found at: http://www.squid-cache.org/Versions/v2/HEAD/changesets/12030.patch * I've modified the delay pool ACL descriptions to point out they work on -requests- and not -replies-. Talk to me privately if you'd like to see -reply- processing for delay pools - it shouldn't take me long to do (I'd guess 10 or 20 hours to code it up, test and commit) so a little sponsorship there would be useful. This will let you do things like placing urls based on reply Content-Type Headers, etc, into different delay pools (ie, what people were trying to do a few weeks ago on squid-users.) Thanks! Adrian -- - Xenion - http://www.xenion.com.au/ - VPS Hosting - Commercial Squid Support - - $25/pm entry-level VPSes w/ capped bandwidth charges available in WA -
