You are viewing a plain text version of this content. The canonical link for it is here.
Posted to cvs@httpd.apache.org by ni...@apache.org on 2010/04/02 00:48:38 UTC
svn commit: r930125 - in /httpd/httpd/trunk: CHANGES
modules/proxy/mod_proxy.c modules/proxy/mod_proxy.h
modules/proxy/mod_proxy_balancer.c
Author: niq
Date: Thu Apr 1 22:48:38 2010
New Revision: 930125
URL: http://svn.apache.org/viewvc?rev=930125&view=rev
Log:
Proxy balancer: support setting error status according to
HTTP response code from a backend.
PR 48939 [Daniel Ruggeri <DRuggeri primary.net>]
Modified:
httpd/httpd/trunk/CHANGES
httpd/httpd/trunk/modules/proxy/mod_proxy.c
httpd/httpd/trunk/modules/proxy/mod_proxy.h
httpd/httpd/trunk/modules/proxy/mod_proxy_balancer.c
Modified: httpd/httpd/trunk/CHANGES
URL: http://svn.apache.org/viewvc/httpd/httpd/trunk/CHANGES?rev=930125&r1=930124&r2=930125&view=diff
==============================================================================
--- httpd/httpd/trunk/CHANGES [utf-8] (original)
+++ httpd/httpd/trunk/CHANGES [utf-8] Thu Apr 1 22:48:38 2010
@@ -28,6 +28,10 @@ Changes with Apache 2.3.7
processing is completed, avoiding orphaned callback pointers.
[Brett Gervasoni <brettg senseofsecurity.com>, Jeff Trawick]
+ *) Proxy balancer: support setting error status according to
+ HTTP response code from a backend.
+ PR 48939 [Daniel Ruggeri <DRuggeri primary.net>]
+
*) htcacheclean: Introduce the ability to clean specific URLs from the
cache, if provided as an optional parameter on the command line.
[Graham Leggett]
Modified: httpd/httpd/trunk/modules/proxy/mod_proxy.c
URL: http://svn.apache.org/viewvc/httpd/httpd/trunk/modules/proxy/mod_proxy.c?rev=930125&r1=930124&r2=930125&view=diff
==============================================================================
--- httpd/httpd/trunk/modules/proxy/mod_proxy.c (original)
+++ httpd/httpd/trunk/modules/proxy/mod_proxy.c Thu Apr 1 22:48:38 2010
@@ -381,6 +381,27 @@ static const char *set_balancer_param(pr
else
return "scolonpathdelim must be On|Off";
}
+ else if (!strcasecmp(key, "erroronstatus")) {
+ char valSplit[strlen(val)+1];
+ char *status;
+ char *tok_state;
+
+ strcpy(valSplit, val);
+ balancer->errstatuses = apr_array_make(p, 1, sizeof(int));
+
+ status = apr_strtok(valSplit, ", ", &tok_state);
+ while (status != NULL) {
+ ival = atoi(status);
+ if (ap_is_HTTP_VALID_RESPONSE(ival)) {
+ *(int*)apr_array_push(balancer->errstatuses) = ival;
+ }
+ else {
+ return "erroronstatus must be one or more HTTP response code";
+ }
+ status = apr_strtok(NULL, ", ", &tok_state);
+ }
+
+ }
else {
return "unknown Balancer parameter";
}
Modified: httpd/httpd/trunk/modules/proxy/mod_proxy.h
URL: http://svn.apache.org/viewvc/httpd/httpd/trunk/modules/proxy/mod_proxy.h?rev=930125&r1=930124&r2=930125&view=diff
==============================================================================
--- httpd/httpd/trunk/modules/proxy/mod_proxy.h (original)
+++ httpd/httpd/trunk/modules/proxy/mod_proxy.h Thu Apr 1 22:48:38 2010
@@ -373,6 +373,7 @@ struct proxy_balancer {
void *context; /* general purpose storage */
const char *sticky_path; /* URL sticky session identifier */
int scolonsep; /* true if ';' seps sticky session paths */
+ apr_array_header_t *errstatuses; /* statuses to force members into error */
};
struct proxy_balancer_method {
Modified: httpd/httpd/trunk/modules/proxy/mod_proxy_balancer.c
URL: http://svn.apache.org/viewvc/httpd/httpd/trunk/modules/proxy/mod_proxy_balancer.c?rev=930125&r1=930124&r2=930125&view=diff
==============================================================================
--- httpd/httpd/trunk/modules/proxy/mod_proxy_balancer.c (original)
+++ httpd/httpd/trunk/modules/proxy/mod_proxy_balancer.c Thu Apr 1 22:48:38 2010
@@ -600,7 +600,6 @@ static int proxy_balancer_post_request(p
proxy_server_conf *conf)
{
-#if 0
apr_status_t rv;
if ((rv = PROXY_THREAD_LOCK(balancer)) != APR_SUCCESS) {
@@ -609,8 +608,20 @@ static int proxy_balancer_post_request(p
balancer->name);
return HTTP_INTERNAL_SERVER_ERROR;
}
- /* TODO: placeholder for post_request actions
- */
+
+ if (!apr_is_empty_array(balancer->errstatuses)){
+ int i;
+ for (i = 0; i < balancer->errstatuses->nelts; i++) {
+ int val=((int*)balancer->errstatuses->elts)[i];
+ if (r->status == val) {
+ ap_log_error(APLOG_MARK, APLOG_NOTICE, rv, r->server,
+ "Detected ErrorOnState (%d) for member (%s). Forcing worker into error state.", val, worker->name);
+ worker->s->status |= PROXY_WORKER_IN_ERROR;
+ worker->s->error_time = apr_time_now();
+ break;
+ }
+ }
+ }
if ((rv = PROXY_THREAD_UNLOCK(balancer)) != APR_SUCCESS) {
ap_log_error(APLOG_MARK, APLOG_ERR, rv, r->server,
@@ -620,8 +631,6 @@ static int proxy_balancer_post_request(p
ap_log_error(APLOG_MARK, APLOG_DEBUG, 0, r->server,
"proxy_balancer_post_request for (%s)", balancer->name);
-#endif
-
if (worker && worker->s->busy)
worker->s->busy--;
Re: svn commit: r930125 - in /httpd/httpd/trunk: CHANGES
modules/proxy/mod_proxy.c modules/proxy/mod_proxy.h modules/proxy/mod_proxy_balancer.c
Posted by Eric Covener <co...@gmail.com>.
>>> Any opinions on "erroronstatus"? It doesn't feel quite right to me,
>>> but I'm not confident about a replacement.
>>>
>>> Just plain "errorstatus" or "errorstatuscodes"?
Tough one without ending up with a full sentence, maybe just settle on
the original and make sure the doc is crisp?
errorstateonstatus
error-state-on-status
(I like disambiguating either the "error" or the "status" with an
extra word, but the length is a downer)
--
Eric Covener
covener@gmail.com
Re: svn commit: r930125 - in /httpd/httpd/trunk: CHANGES modules/proxy/mod_proxy.c
modules/proxy/mod_proxy.h modules/proxy/mod_proxy_balancer.c
Posted by Daniel Ruggeri <DR...@primary.net>.
On 8/19/2010 7:57 PM, Jeff Trawick wrote:
> On Tue, Jul 20, 2010 at 10:59 AM, Daniel Ruggeri <DRuggeri@primary.net
> <ma...@primary.net>> wrote:
>
> On 7/16/2010 10:37 AM, Jeff Trawick wrote:
>
> On Fri, Jul 16, 2010 at 11:27 AM, William A. Rowe Jr.
> <wrowe@rowe-clan.net <ma...@rowe-clan.net>> wrote:
>
> On 7/16/2010 9:35 AM, William A. Rowe Jr. wrote:
>
> On 7/16/2010 6:47 AM, Jeff Trawick wrote:
>
> We may as well leave it at "erroronstatus" I
> agree with Eric on
> keeping your original choice and worrying about
> the doc.
>
> Can we drop the word error, since it has four
> potential meanings?
>
> backendfailureonstatus perhaps, or something along
> those lines?
>
> Hold up; see
> http://tomcat.apache.org/connectors-doc/reference/workers.html
> fail_on_status - why would we invent new phrases?
>
> good catch
>
>
>
> Sounds like this would be the best name. As promised, here is the
> final patch including doc, the name "failonstatus", and updates
> made in trunk for this patch. Also included is an update to trunk
> for the new name and doc.
>
>
> Thanks; I added one tweak to the patch -- an "s" at the end of "code"
> in the following line:
>
> + return "erroronstatus must be one or more HTTP response code";
>
> + }
>
>
> For ease of reading, the doc patch reads thusly:
> A single or comma-separated list of HTTP status codes. If
> set this will
> force the worker into error state when the backend returns
> any status code
> in the list. Worker recovery behaves the same as other
> worker errors.
>
> Finally, here is my proposed change to STATUS to indicate
> completion of Jeff's suggestions as well as update the location of
> the 2.2.x patch:
> --- STATUS 2010-07-20 09:53:36.298789200 -0500
> +++ STATUS.mod 2010-07-20 09:54:15.466286400 -0500
> @@ -145,13 +145,9 @@
> statuses are found
> PR: 48939
> Trunk patch: http://svn.apache.org/viewvc?rev=930125&view=rev
> <http://svn.apache.org/viewvc?rev=930125&view=rev>
> - 2.2.x patch:
> https://issues.apache.org/bugzilla/attachment.cgi?id=25153
> + 2.2.x patch:
> https://issues.apache.org/bugzilla/attachment.cgi?id=25788
>
>
> Any chance you can update your 2.2.x patch in Bugzilla to include the
> tweaked wording (s/code/codes/)? Sorry :)
>
> Submitted by: Daniel Ruggeri <DRuggeri primary.net
> <http://primary.net>>
>
> +1: niq, jim
> - trawick suggests:
> - 1. somebody write doc (Daniel volunteers)
> - 2. somebody create new patch which includes r962972, any
> - subsequent changes, and doc (Daniel volunteers)
>
> * mod_disk_cache: Decline the opportunity to cache if the
> response is
> a 206 Partial Content. This stops a reverse proxied partial
> response
>
>
> Thank you, all.
> --
> Daniel Ruggeri
>
>
Jeff;
I'm not sure what happened - the 2.2 patch I submitted had a mixture
between "failonstatus" and "erroronstatus", but the trunk patch did not.
I only noticed because of the message you posted about patching the
patch and the update didn't seem right. In any event, I have updated
Bugzilla and the new 2.2 patch can be found here:
https://issues.apache.org/bugzilla/attachment.cgi?id=25923
Sorry about the confusion.
Status update should look like this (new attachment ID included)
--- STATUS 2010-07-20 09:53:36.298789200 -0500
+++ STATUS.mod 2010-07-20 09:54:15.466286400 -0500
@@ -145,13 +145,9 @@
statuses are found
PR: 48939
Trunk patch: http://svn.apache.org/viewvc?rev=930125&view=rev
- 2.2.x patch: https://issues.apache.org/bugzilla/attachment.cgi?id=25153
+ 2.2.x patch: https://issues.apache.org/bugzilla/attachment.cgi?id=25923
Submitted by: Daniel Ruggeri<DRuggeri primary.net>
+1: niq, jim
- trawick suggests:
- 1. somebody write doc (Daniel volunteers)
- 2. somebody create new patch which includes r962972, any
- subsequent changes, and doc (Daniel volunteers)
* mod_disk_cache: Decline the opportunity to cache if the response is
a 206 Partial Content. This stops a reverse proxied partial response
Re: svn commit: r930125 - in /httpd/httpd/trunk: CHANGES
modules/proxy/mod_proxy.c modules/proxy/mod_proxy.h modules/proxy/mod_proxy_balancer.c
Posted by Jeff Trawick <tr...@gmail.com>.
On Tue, Jul 20, 2010 at 10:59 AM, Daniel Ruggeri <DR...@primary.net>wrote:
> On 7/16/2010 10:37 AM, Jeff Trawick wrote:
>
>> On Fri, Jul 16, 2010 at 11:27 AM, William A. Rowe Jr.
>> <wr...@rowe-clan.net> wrote:
>>
>>
>>> On 7/16/2010 9:35 AM, William A. Rowe Jr. wrote:
>>>
>>>
>>>> On 7/16/2010 6:47 AM, Jeff Trawick wrote:
>>>>
>>>>
>>>>> We may as well leave it at "erroronstatus" I agree with Eric on
>>>>> keeping your original choice and worrying about the doc.
>>>>>
>>>>>
>>>> Can we drop the word error, since it has four potential meanings?
>>>>
>>>> backendfailureonstatus perhaps, or something along those lines?
>>>>
>>>>
>>> Hold up; see
>>> http://tomcat.apache.org/connectors-doc/reference/workers.html
>>> fail_on_status - why would we invent new phrases?
>>>
>>>
>> good catch
>>
>>
>>
>>
> Sounds like this would be the best name. As promised, here is the final
> patch including doc, the name "failonstatus", and updates made in trunk for
> this patch. Also included is an update to trunk for the new name and doc.
>
Thanks; I added one tweak to the patch -- an "s" at the end of "code" in the
following line:
+ return "erroronstatus must be one or more HTTP response code";
+ }
>
> For ease of reading, the doc patch reads thusly:
> A single or comma-separated list of HTTP status codes. If set this
> will
> force the worker into error state when the backend returns any
> status code
> in the list. Worker recovery behaves the same as other worker
> errors.
>
> Finally, here is my proposed change to STATUS to indicate completion of
> Jeff's suggestions as well as update the location of the 2.2.x patch:
> --- STATUS 2010-07-20 09:53:36.298789200 -0500
> +++ STATUS.mod 2010-07-20 09:54:15.466286400 -0500
> @@ -145,13 +145,9 @@
> statuses are found
> PR: 48939
> Trunk patch: http://svn.apache.org/viewvc?rev=930125&view=rev
> - 2.2.x patch:
> https://issues.apache.org/bugzilla/attachment.cgi?id=25153
> + 2.2.x patch:
> https://issues.apache.org/bugzilla/attachment.cgi?id=25788
>
Any chance you can update your 2.2.x patch in Bugzilla to include the
tweaked wording (s/code/codes/)? Sorry :)
> Submitted by: Daniel Ruggeri <DRuggeri primary.net>
>
> +1: niq, jim
> - trawick suggests:
> - 1. somebody write doc (Daniel volunteers)
> - 2. somebody create new patch which includes r962972, any
> - subsequent changes, and doc (Daniel volunteers)
>
> * mod_disk_cache: Decline the opportunity to cache if the response is
> a 206 Partial Content. This stops a reverse proxied partial response
>
>
> Thank you, all.
> --
> Daniel Ruggeri
>
--
Born in Roswell... married an alien...
Re: svn commit: r930125 - in /httpd/httpd/trunk: CHANGES modules/proxy/mod_proxy.c
modules/proxy/mod_proxy.h modules/proxy/mod_proxy_balancer.c
Posted by Daniel Ruggeri <DR...@primary.net>.
On 7/16/2010 10:37 AM, Jeff Trawick wrote:
> On Fri, Jul 16, 2010 at 11:27 AM, William A. Rowe Jr.
> <wr...@rowe-clan.net> wrote:
>
>> On 7/16/2010 9:35 AM, William A. Rowe Jr. wrote:
>>
>>> On 7/16/2010 6:47 AM, Jeff Trawick wrote:
>>>
>>>> We may as well leave it at "erroronstatus" I agree with Eric on
>>>> keeping your original choice and worrying about the doc.
>>>>
>>> Can we drop the word error, since it has four potential meanings?
>>>
>>> backendfailureonstatus perhaps, or something along those lines?
>>>
>> Hold up; see http://tomcat.apache.org/connectors-doc/reference/workers.html
>> fail_on_status - why would we invent new phrases?
>>
> good catch
>
>
>
Sounds like this would be the best name. As promised, here is the final
patch including doc, the name "failonstatus", and updates made in trunk
for this patch. Also included is an update to trunk for the new name and
doc.
For ease of reading, the doc patch reads thusly:
A single or comma-separated list of HTTP status codes. If set
this will
force the worker into error state when the backend returns any
status code
in the list. Worker recovery behaves the same as other worker
errors.
Finally, here is my proposed change to STATUS to indicate completion of
Jeff's suggestions as well as update the location of the 2.2.x patch:
--- STATUS 2010-07-20 09:53:36.298789200 -0500
+++ STATUS.mod 2010-07-20 09:54:15.466286400 -0500
@@ -145,13 +145,9 @@
statuses are found
PR: 48939
Trunk patch: http://svn.apache.org/viewvc?rev=930125&view=rev
- 2.2.x patch: https://issues.apache.org/bugzilla/attachment.cgi?id=25153
+ 2.2.x patch: https://issues.apache.org/bugzilla/attachment.cgi?id=25788
Submitted by: Daniel Ruggeri <DRuggeri primary.net>
+1: niq, jim
- trawick suggests:
- 1. somebody write doc (Daniel volunteers)
- 2. somebody create new patch which includes r962972, any
- subsequent changes, and doc (Daniel volunteers)
* mod_disk_cache: Decline the opportunity to cache if the response is
a 206 Partial Content. This stops a reverse proxied partial response
Thank you, all.
--
Daniel Ruggeri
Re: svn commit: r930125 - in /httpd/httpd/trunk: CHANGES
modules/proxy/mod_proxy.c modules/proxy/mod_proxy.h modules/proxy/mod_proxy_balancer.c
Posted by Jeff Trawick <tr...@gmail.com>.
On Fri, Jul 16, 2010 at 11:27 AM, William A. Rowe Jr.
<wr...@rowe-clan.net> wrote:
> On 7/16/2010 9:35 AM, William A. Rowe Jr. wrote:
>> On 7/16/2010 6:47 AM, Jeff Trawick wrote:
>>>
>>> We may as well leave it at "erroronstatus" I agree with Eric on
>>> keeping your original choice and worrying about the doc.
>>
>> Can we drop the word error, since it has four potential meanings?
>>
>> backendfailureonstatus perhaps, or something along those lines?
>
> Hold up; see http://tomcat.apache.org/connectors-doc/reference/workers.html
> fail_on_status - why would we invent new phrases?
good catch
Re: svn commit: r930125 - in /httpd/httpd/trunk: CHANGES modules/proxy/mod_proxy.c
modules/proxy/mod_proxy.h modules/proxy/mod_proxy_balancer.c
Posted by "William A. Rowe Jr." <wr...@rowe-clan.net>.
On 7/16/2010 9:35 AM, William A. Rowe Jr. wrote:
> On 7/16/2010 6:47 AM, Jeff Trawick wrote:
>>
>> We may as well leave it at "erroronstatus" I agree with Eric on
>> keeping your original choice and worrying about the doc.
>
> Can we drop the word error, since it has four potential meanings?
>
> backendfailureonstatus perhaps, or something along those lines?
Hold up; see http://tomcat.apache.org/connectors-doc/reference/workers.html
fail_on_status - why would we invent new phrases?
Re: svn commit: r930125 - in /httpd/httpd/trunk: CHANGES modules/proxy/mod_proxy.c
modules/proxy/mod_proxy.h modules/proxy/mod_proxy_balancer.c
Posted by "William A. Rowe Jr." <wr...@rowe-clan.net>.
On 7/16/2010 6:47 AM, Jeff Trawick wrote:
>
> We may as well leave it at "erroronstatus" I agree with Eric on
> keeping your original choice and worrying about the doc.
Can we drop the word error, since it has four potential meanings?
backendfailureonstatus perhaps, or something along those lines?
Re: svn commit: r930125 - in /httpd/httpd/trunk: CHANGES
modules/proxy/mod_proxy.c modules/proxy/mod_proxy.h modules/proxy/mod_proxy_balancer.c
Posted by Jeff Trawick <tr...@gmail.com>.
On Thu, Jul 15, 2010 at 8:36 PM, Daniel Ruggeri <DR...@primary.net> wrote:
> On 7/12/2010 2:17 PM, Jim Jagielski wrote:
>>
>> On Jul 11, 2010, at 11:04 AM, Jeff Trawick wrote:
>>
>>
>>>
>>> On Thu, Apr 1, 2010 at 6:48 PM,<ni...@apache.org> wrote:
>>>
>>>>
>>>> Author: niq
>>>> Date: Thu Apr 1 22:48:38 2010
>>>> New Revision: 930125
>>>>
>>>> URL: http://svn.apache.org/viewvc?rev=930125&view=rev
>>>> Log:
>>>> Proxy balancer: support setting error status according to
>>>> HTTP response code from a backend.
>>>> PR 48939 [Daniel Ruggeri<DRuggeri primary.net>]
>>>>
>>>
>>>
>>>>
>>>> + else if (!strcasecmp(key, "erroronstatus")) {
>>>>
>>>
>>> Any opinions on "erroronstatus"? It doesn't feel quite right to me,
>>> but I'm not confident about a replacement.
>>>
>>> Just plain "errorstatus" or "errorstatuscodes"?
>>>
>>> If normal response handling takes the same action on other status
>>> codes (I dunno), maybe "extraerrorstatus" or "extraerrorstatuscodes"
>>> would be bettter?
>>>
>>>
>>
>> proxyseterror?
>>
>>
>>
>
> If erroronstatus is right out, I like errorstatuscodes best as a
> replacement.
>
> Jeff;
> I will gladly take the tasks you mentioned in STATUS as soon as we have
> consensus on the name of the directive. It may be a lofty goal but it would
> be great if this could make it into 2.2.16.
We may as well leave it at "erroronstatus" I agree with Eric on
keeping your original choice and worrying about the doc.
Re: svn commit: r930125 - in /httpd/httpd/trunk: CHANGES modules/proxy/mod_proxy.c
modules/proxy/mod_proxy.h modules/proxy/mod_proxy_balancer.c
Posted by Daniel Ruggeri <DR...@primary.net>.
On 7/12/2010 2:17 PM, Jim Jagielski wrote:
> On Jul 11, 2010, at 11:04 AM, Jeff Trawick wrote:
>
>
>> On Thu, Apr 1, 2010 at 6:48 PM,<ni...@apache.org> wrote:
>>
>>> Author: niq
>>> Date: Thu Apr 1 22:48:38 2010
>>> New Revision: 930125
>>>
>>> URL: http://svn.apache.org/viewvc?rev=930125&view=rev
>>> Log:
>>> Proxy balancer: support setting error status according to
>>> HTTP response code from a backend.
>>> PR 48939 [Daniel Ruggeri<DRuggeri primary.net>]
>>>
>>
>>> + else if (!strcasecmp(key, "erroronstatus")) {
>>>
>> Any opinions on "erroronstatus"? It doesn't feel quite right to me,
>> but I'm not confident about a replacement.
>>
>> Just plain "errorstatus" or "errorstatuscodes"?
>>
>> If normal response handling takes the same action on other status
>> codes (I dunno), maybe "extraerrorstatus" or "extraerrorstatuscodes"
>> would be bettter?
>>
>>
> proxyseterror?
>
>
>
If erroronstatus is right out, I like errorstatuscodes best as a
replacement.
Jeff;
I will gladly take the tasks you mentioned in STATUS as soon as we
have consensus on the name of the directive. It may be a lofty goal but
it would be great if this could make it into 2.2.16.
--
Daniel Ruggeri
Re: svn commit: r930125 - in /httpd/httpd/trunk: CHANGES modules/proxy/mod_proxy.c modules/proxy/mod_proxy.h modules/proxy/mod_proxy_balancer.c
Posted by Jim Jagielski <ji...@jaguNET.com>.
On Jul 11, 2010, at 11:04 AM, Jeff Trawick wrote:
> On Thu, Apr 1, 2010 at 6:48 PM, <ni...@apache.org> wrote:
>> Author: niq
>> Date: Thu Apr 1 22:48:38 2010
>> New Revision: 930125
>>
>> URL: http://svn.apache.org/viewvc?rev=930125&view=rev
>> Log:
>> Proxy balancer: support setting error status according to
>> HTTP response code from a backend.
>> PR 48939 [Daniel Ruggeri <DRuggeri primary.net>]
>
>> + else if (!strcasecmp(key, "erroronstatus")) {
>
> Any opinions on "erroronstatus"? It doesn't feel quite right to me,
> but I'm not confident about a replacement.
>
> Just plain "errorstatus" or "errorstatuscodes"?
>
> If normal response handling takes the same action on other status
> codes (I dunno), maybe "extraerrorstatus" or "extraerrorstatuscodes"
> would be bettter?
>
proxyseterror?
Re: svn commit: r930125 - in /httpd/httpd/trunk: CHANGES
modules/proxy/mod_proxy.c modules/proxy/mod_proxy.h modules/proxy/mod_proxy_balancer.c
Posted by Jeff Trawick <tr...@gmail.com>.
On Thu, Apr 1, 2010 at 6:48 PM, <ni...@apache.org> wrote:
> Author: niq
> Date: Thu Apr 1 22:48:38 2010
> New Revision: 930125
>
> URL: http://svn.apache.org/viewvc?rev=930125&view=rev
> Log:
> Proxy balancer: support setting error status according to
> HTTP response code from a backend.
> PR 48939 [Daniel Ruggeri <DRuggeri primary.net>]
> + else if (!strcasecmp(key, "erroronstatus")) {
Any opinions on "erroronstatus"? It doesn't feel quite right to me,
but I'm not confident about a replacement.
Just plain "errorstatus" or "errorstatuscodes"?
If normal response handling takes the same action on other status
codes (I dunno), maybe "extraerrorstatus" or "extraerrorstatuscodes"
would be bettter?