If your default_ttl is not 0, then this may be the expected behavior. I'm not
sure if Varnish should really ever cache >=500 responses?
But in VCL you could do something like:
sub vcl_fetch {
if ( obj.status >= 500 ) {
set obj.ttl = 0s;
set obj.cacheable = false;
}
}
Adjusting timeouts for ttl for 404s is also handy. Hope it helps,
--
Ken
On Feb 2, 2010, at 1:37 PM, pub crawler wrote:
> Thought I'd ask the list before I went on a voyage with this one.
>
> Sometime our backend app servers gets overloaded and go into
> protection mode whereby they sends out 503 errors until they recover.
>
> Varnish is in front of the app servers and when this happens the 503
> ends up as a cached item in Varnish.
>
> Here's output to show such:
>
> HTTP/1.1 503 Service Unavailable
> Keep-Alive: timeout=30
> Content-Type: text/html
> Content-Language: en-US
> Content-Length: 122
> Date: Tue, 02 Feb 2010 21:34:15 GMT
> X-Varnish: 655791101 655790499
> Age: 97
> Via: 1.1 varnish
> Connection: keep-alive
> X-Served-By: atom3302
> X-Cache: HIT
> X-Cache-Hits: 13
>
> What is the best way to explicitly refuse caching of 503 errors by Varnish?
>
> Thanks!
> _______________________________________________
> varnish-misc mailing list
> varnish-misc@projects.linpro.no
> http://projects.linpro.no/mailman/listinfo/varnish-misc
_______________________________________________
varnish-misc mailing list
varnish-misc@projects.linpro.no
http://projects.linpro.no/mailman/listinfo/varnish-misc
