Hello,
I've got one odditiy with Squid 2.1/P1 which I would label bug and a
wishlist item.
First the bug, today one user downloaded a 60MB file (the StarOffice
5.0 release in case you're wondering ;-) via one of the 3 similar
configured proxies here. The file was successfully cached, I upgraded
to 2.1 (and wasted 6 GB of perfectly healthy V1.22 cached objects) among
other things to better support such massive files. When trying to access
the same (FTP) URL from another sibling cache here it started a _DIRECT_
download instead of using the cached copy in it's neighbor, even though
this neighbor recorded the following UDP_HIT:
--- 911749333.979 0 nausicaa.swb.de UDP_HIT/000 89 ICP_QUERY ftp://ftp.tu-clausthal.de/pub/linux/StarDivision/wntmsci/so50_49.exe - NONE/- - --- I retried it with the sibling relation replaced by a parent one, no difference... Nothing in the configurations (AFAIK) would prevent the usage of the cached copy from the neighbor (max object size is at 96MB), what could be the reason for this behavior? One really nice thing to have would be a cache-manager option to abort a specific download (w/o shutting down the complete Squid). In the example above the user was downloading the 60MB from 2 mirrors in parallel, one of them was a tad slower, so I wanted to stop this from wasting bandwidth. I finally did a "route add <hostip> gw 127.0.0.4" and left it to time out, but that's not what I would call elegant... ^_^ Dewa, <CB> -- // <CB> aka Christian Balzer, Tannenstr. 23c, D-64342 Seeheim, Germany \X/ CB@aichan.swb.de | Voice: +49 6257 83036, Fax/Data: +49 6257 83037 SWB - The Software Brewery - | http://www.swb.de/ | Anime no OtakuReceived on Sun Nov 22 1998 - 12:36:56 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:43:12 MST