My bad, I didn't figure out you were talking about 503s from the
backend (as opposed to eg. connect errors which are handled as
internal 503s).
AFAICT recovering from backend error is not reliably feasible, non
idempotent requests (eg. POSTs) cannot be re-sent by a proxy.
On Fri, May 15, 2015 at 11:53 AM, nik600 <nik600@xxxxxxxxx> wrote:
> hhmmm i'm not experiencing that.
>
> i've got a php script on the back end that send a 503 header:
>
> NODE1
> <?php
> header('HTTP/1.1 503 Service unavailable');
> echo "error";
> ?>
>
> NODE2
> <?php
> echo "ok";
> ?>
>
> But when i make requests on the cluster i get 50% of "error" and 50% of
> "ok".
>
>
>
> 2015-05-15 11:09 GMT+02:00 Yann Ylavic <ylavic.dev@xxxxxxxxx>:
>>
>> Hi,
>>
>> this is an implicit behaviour.
>>
>> Regards,
>> Yann.
>>
>> On Fri, May 15, 2015 at 10:31 AM, nik600 <nik600@xxxxxxxxx> wrote:
>> > Dear all
>> >
>> > is possible to handle the specific http 503 status from a BalancerMember
>> > and
>> > re-forward the request to another member?
>> >
>> > Thanks to all in advance
>> >
>> > --
>> > /*************/
>> > nik600
>> > http://www.kumbe.it
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx
>> For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx
>>
>
>
>
> --
> /*************/
> nik600
> http://www.kumbe.it
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@xxxxxxxxxxxxxxxx
For additional commands, e-mail: users-help@xxxxxxxxxxxxxxxx
|