Re: [PATCH] mod_proxy run cleanup on balancer failure
Brian Akins wrote:
> Can we get a vote on this? Jim gave it +1. So do I. I would like this
> to be in a beta "release." This fix is important to our environment.
Never mind. See this in trunk...
--
Brian Akins
Lead Systems Engineer
CNN Internet Technologies
Here is a trivial patch that will allow proxy_handler to run the request_status hook if pre_request fails. This is necessary if all balncer members are in an...
Brian Akins
bakins@...
Sep 28, 2005 2:29 pm
... If all workers are in error state, worker will be NULL. Are you sure that the post_request hook that is run in the balancer case is prepared for such a...
r.pluem@...
Sep 28, 2005 9:09 pm
... Hmmm... It might be best to, in ap_proxy_pre_request() set *balancer to NULL if the result from proxy_run_pre_request() isn't OK. Then we can call jump to...
Jim Jagielski
jim@...
Sep 29, 2005 12:59 pm
... [..cut..] ... Yes, this sounds like the better solution here. ... Yes, I think you are right. It should be *balancer. Regards Rüdiger...
r.pluem@...
Sep 29, 2005 1:10 pm
... goto cleanup; will actually run the request_status optional hook, so the patch makes sense, although I would make it as follows: if (access_status != OK) {...
Mladen Turk
mturk@...
Sep 29, 2005 7:26 am
... I guess that makes sense. I just want to chatch the following cases: No available workers - all in error or busy worker failed - ie connect timeout random...
Brian Akins
brian.akins@...
Sep 29, 2005 1:41 pm
... The rub, as mentioned, is to ensure that when jumping to cleanup, balancer is never non-NULL with a NULL worker, so we need to check if that ever happens....
Jim Jagielski
jim@...
Sep 29, 2005 2:49 pm
... What about the following patch? I think it should address all the things discussed. Regards Rüdiger Index: mod_proxy.c ... +++ mod_proxy.c (Arbeitskopie) ...
r.pluem@...
Sep 29, 2005 4:01 pm
Except for the formatting +1 :)...
Jim Jagielski
jim@...
Sep 29, 2005 4:43 pm
As I try to improve my Apache code style awareness. What is wrong with the formatting? Regards Rüdiger...
r.pluem@...
Sep 29, 2005 6:33 pm
Sorry :) The single line if statements. The pref is if (foo) banana(); rather than if (foo) banana();...
Jim Jagielski
jim@...
Sep 29, 2005 6:50 pm
... Thanks. So I guess this should be fine: Index: mod_proxy.c ... +++ mod_proxy.c (Arbeitskopie) @@ -679,8 +679,22 @@ char *url = uri; /* Try to obtain the...
r.pluem@...
Sep 29, 2005 7:49 pm
Can we get a vote on this? Jim gave it +1. So do I. I would like this to be in a beta "release." This fix is important to our environment. ... -- Brian...
Brian Akins
bakins@...
Oct 28, 2005 1:45 pm
... Never mind. See this in trunk... -- Brian Akins Lead Systems Engineer CNN Internet Technologies...
Brian Akins
brian.akins@...
Oct 28, 2005 1:47 pm
... It is also backported to 2.2.x. So it will be part of 2.1.9 Regards Rüdiger...
Ruediger Pluem
rpluem@...
Oct 28, 2005 5:14 pm
... http://httpd.apache.org/dev/styleguide.html The basic objection to your patch would be that statements after your if's; put them on the next line. HTH, ...
Sander Striker
s.striker@...
Sep 29, 2005 7:32 pm
... This means if access_status is DECLINED, we would return. Would that work for you??...
Jim Jagielski
jim@...
Sep 29, 2005 4:31 pm
... Yes. Just need to run request_status in case of error. -- Brian Akins Lead Systems Engineer CNN Internet Technologies...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
Send Email