Re: Patch to make Reload into IMS work more sanely

Stewart Forster writes:

>Hiya,
>
> Please tell me if I'm correct....
>
> When doing that patch I discovered that the logging for TCP_IMS_HIT
>and TCP_IMS_MISS is actually referring to the requesting client's nature
>of a HIT or MISS, and not the responding cache.
>
>That is:
>
>client sends an IMS request with a last-modified date.
>cache checks last-modified against date of copy it has.
>
>Roughly (assuming cache's object is not stale):
>
>If client last modified > cache last modified then
> cache returns 304 (Not Modified)
> cache logs TCP_IMS_HIT
>else
> cache returns 200 and a copy of the object
> cache logs TCP_IMS_MISS
>fi
>
>So an IMS request from a client for a non-stale object in the cache results
>in a hit for that cache either way, but the log text says a MISS.

Yeah, its misleading.

How about if we change

        TCP_IMS_HIT --> TCP_IMS_NM
        TCP_IMS_MISS --> TCP_IMS_HIT

Duane W.
Received on Tue Jul 29 2003 - 13:15:51 MDT