[squid-users] TCP_MISS only

Amos Jeffries squid3 at treenet.co.nz
Tue Sep 27 18:43:35 UTC 2022


On 27/09/22 23:01, Andy Armstrong wrote:
> Hi Alex,
> 
> That makes a lot of sense, I don’t know how I overlooked that – thank 
> you. I also agree, logically caching a 201 response makes little sense, 
> and it was just an example I had that was easy to try so I used that.
> 
> I just altered the HTTP Return code so it sent 200 instead of 201, and 
> the result is sadly the same, I get many, many lines like this:
> 

Unfortunately that is not enough. POST method is also not cacheable by 
default. See <https://www.rfc-editor.org/rfc/rfc9110#section-9.3.3>

Consider what would happen when two clients POST different sets of data 
to the same URL.  Which one should the cache handle *instead* of letting 
it be delivered to a server?


> 1664272638.44310107 10.1.1.70 TCP_MISS/200 275 POST 
> http://192.168.0.2:3001/InternalCommunicationServices/message/email - 
> HIER_DIRECT/192.168.0.2 application/json
> 
> My suspicion is still that my refresh_pattern is wrong:
> 
> refresh_pattern -i http:\/\/129.168.0.2:3001\/.* 10080 100% 43200 
> override-lastmod
> 

refresh_pattern directive does not make things cacheable when they are 
not. It can only extend or shrink cacheability times.


HTH
Amos


More information about the squid-users mailing list