On Monday 18 November 96, at 10 h 33, the keyboard of Richard Munro
<munro@moe.acad.humberc.on.ca> wrote:
> Obviously if an IMS request produces newer material, the old material will
> be replaced on disk with the newer material, which starts up the cycle
> again. But what happens if an IMS request finds that there has been no
> motification? Does the expiry date get modified in the cache or do all
> subsequent client requests go through the IMS process again and again
> until there actually is a modification of the source material?
It seems that Squid 1.1beta19 does not update the "age" of the object
(entry->timestamp, the length of its stay in the cache) when receiving a
"Not modified" so it asks over and over the same object because (use
"debug_options 22,3" to see refreshes in cache.log) "refreshCheck: YES:
factor > pc":
848334708.231 3817 157.99.60.23 TCP_REFRESH_HIT/200 199 GET
http://www.globenet.org/tmp/bidon3 - DIRECT/www.globenet.org
848334710.026 759 157.99.60.23 TCP_REFRESH_HIT/200 199 GET
http://www.globenet.org/tmp/bidon3 - DIRECT/www.globenet.org
848334714.720 3661 157.99.60.23 TCP_REFRESH_HIT/200 199 GET
http://www.globenet.org/tmp/bidon3 - DIRECT/www.globenet.org
848334726.955 11210 157.99.60.23 TCP_REFRESH_HIT/200 199 GET
http://www.globenet.org/tmp/bidon3 - DIRECT/www.globenet.org
Isn't it a bug?
Received on Mon Nov 18 1996 - 08:37:18 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:33:34 MST