Hi Frederik,
On Mon, Mar 28, 2016 at 02:31:27PM -0700, Frederik Deweerdt wrote:
> Hi,
>
> I've been working on an issue we've been seeing on very high loads with
> a configuration that boils down to:
>
> [SSL Traffic] ---> [HAProxy] ---[via send_proxy]--> [Proxy]
>
> We're seeing this with 1.
Hi Chris,
On Thu, Mar 17, 2016 at 11:38:56PM +, Chris Warren wrote:
> We???re currently testing a patch which adds an ???initial-state up/down???
> option to each server (and the default-server option) - the default behaviour
> remains unchanged:
> https://github.com/beamly/haproxy-1.6/commit/
On Tue, Mar 29, 2016 at 06:06:50PM -0600, Shawn Heisey wrote:
> You can only set up a redirect on a different frontend, listening on
> another port WITHOUT SSL. And to do that, I would put the redirect in
> the frontend, not the backend.
>
> Here's a slightly redacted example of what I'm saying:
On Tue, Mar 29, 2016 at 7:21 PM, Paul Draper wrote:
> As I understand it, there seems to no way to add a server to HAProxy without
> incurring significant disruption. Adding a server requires reloading
> configuration. This loses all statistics, all health check information, etc.
> So, for instanc
On Sat, Mar 26, 2016 at 7:16 PM, Vincent Bernat wrote:
> ❦ 26 mars 2016 23:40 +0600, Александр Лебедев
> :
>
>> Hi, Vincent! Thanks for your answer!
>> I made some tests today. Yes, it crashes only if hostname contains an
>> odd number of symbols!
>
> So, it should be easy to fix. Baptiste, do
On 3/29/2016 4:56 PM, Colin Leavett-Brown wrote:
> I have the following haproxy configuration:
>
> global
> daemon
> maxconn 2048
> tune.ssl.default-dh-param 1024
>
> defaults
> mode http
> timeout connect 5000ms
> timeout client 5ms
> timeout
I have the following haproxy configuration:
global
daemon
maxconn 2048
tune.ssl.default-dh-param 1024
defaults
mode http
timeout connect 5000ms
timeout client 5ms
timeout server 5ms
frontend keystone_public
bind beaver.heprc.uvic.
On Tue, Mar 29, 2016 at 01:04:44PM -0700, Maciej Katafiasz wrote:
> On 29 March 2016 at 07:52, Willy Tarreau wrote:
> > On Sun, Mar 27, 2016 at 05:41:08PM +0200, Vincent Bernat wrote:
> >> Willy, you always cast for functions returning void*. This is not
> >> needed. For example:
> >>
> >> l = (st
On 29 March 2016 at 07:52, Willy Tarreau wrote:
> On Sun, Mar 27, 2016 at 05:41:08PM +0200, Vincent Bernat wrote:
>> Willy, you always cast for functions returning void*. This is not
>> needed. For example:
>>
>> l = (struct listener *)calloc(1, sizeof(struct listener));
>>
>> Could be just:
>>
>>
(The example should have been s3-1.amazonaws.com.)
On Tue, Mar 29, 2016 at 11:21 AM, Paul Draper wrote:
> As I understand it, there seems to no way to add a server to HAProxy
> without incurring significant disruption. Adding a server requires
> reloading configuration. This loses all statistics
> On 29 Mar 2016, at 18:22, Paul Draper wrote:
>
> As I understand it, there seems to no way to add a server to HAProxy without
> incurring significant disruption. Adding a server requires reloading
> configuration. This loses all statistics, all health check information, etc.
> So, for inst
中国国际汽车零部件博览会
China International Auto Parts Expo
中华人民共和国商务部引导支持展会
国家级国际性汽车配件用品展贸平台
【中文名称】 2016第十四届中国(广州)国际汽车零部件及用品展览会
【英文名称】 The 14th China (Guangzhou) International Auto Parts Expo,2016 (CAPE 2016)
【展会日期】 2016年06月08—10日
【展会场馆】 广州琶洲保利世贸博览馆
【展会简介】
本届CAPE预计展会面积67000平方米,标准展位3000多个,参观观众65330多人
As I understand it, there seems to no way to add a server to HAProxy
without incurring significant disruption. Adding a server requires
reloading configuration. This loses all statistics, all health check
information, etc. So, for instance, after adding a server, HAProxy will
send traffic to unheal
nks a lot for the fast investigation! The proposed patch seems to
do the trick :)
Hrm, or not. At least not completely.
There's still something wrong it seems:
20160329 15:07:03: 0x3bca858: key=xx.xx.xx.xx use=0 exp=28799601
gpc0=0 conn_cnt=682 conn_rate(1)=1 conn_cur=3 sess_cnt
On Tue, Mar 29, 2016 at 05:44:17PM +0200, Vincent Bernat wrote:
> ??? 29 mars 2016 17:27 +0200, Willy Tarreau :
>
> >>
> >> @@
> >> type T;
> >> @@
> >>
> >> - (T\( \|\)*)
> >> (\(lua_touserdata\|malloc\|calloc\)(...))
> >>
> >> So, I can rebase the patch as long as it's needed.
> >
> > Per
❦ 29 mars 2016 17:27 +0200, Willy Tarreau :
>>
>> @@
>> type T;
>> @@
>>
>> - (T\( \|\)*)
>> (\(lua_touserdata\|malloc\|calloc\)(...))
>>
>> So, I can rebase the patch as long as it's needed.
>
> Perfect. Then I'll try to flush the large queue ASAP so that we can
> apply such changes. If yo
On Tue, Mar 29, 2016 at 05:03:23PM +0200, Vincent Bernat wrote:
> OK, so, I'll propose a patch for this, including the sizeof() change (or
> another patch for the sizeof() change?). For the record, I don't do this
> manually, I use a semantic patch like this:
>
> @@
> type T;
> @@
>
> - (T\( \|\)
❦ 29 mars 2016 16:52 +0200, Willy Tarreau :
>l = calloc(1, sizeof(*l));
[...]
>> I can propose a (large) patch to remove all those casts (using a
>> semantic patch to not miss anything). Here is a preview (for master,
>> only src/):
>>
>> https://gist.github.com/dc61d8b035545dc24efd
>>
>>
Hi Vincent,
On Sun, Mar 27, 2016 at 05:41:08PM +0200, Vincent Bernat wrote:
> I am trying to look if there are other cases of alignment problem. Some
> time ago, gcc was emitting a warning if this was the case (-Wcast-align)
> but there was too many false positives and this is not the case anymore
Hi Daniel,
On Tue, Mar 29, 2016 at 02:16:55PM +0200, Daniel Schneller wrote:
> Hi!
>
> I am seeing a segfault upon the first request coming through the
> configuration below.
>
> My intention is to enforce a) a total request limit per minute and b) a
> separate limit for certain API paths. For
On Tue, Mar 29, 2016 at 01:14:30PM +0200, Nenad Merdanovic wrote:
> Similar issue was fixed in 67dad27, but the fix is incomplete. Crash still
> happened when utilizing req.fhdr() and sending exactly MAX_HDR_HISTORY
> headers.
>
> This fix needs to be backported to 1.5 and 1.6.
Thanks Nenad for f
ch seems to
do the trick :)
Hrm, or not. At least not completely.
There's still something wrong it seems:
20160329 15:07:03: 0x3bca858: key=xx.xx.xx.xx use=0 exp=28799601 gpc0=0
conn_cnt=682 conn_rate(1)=1 conn_cur=3 sess_cnt=1
sess_rate(1)=-1032058827 http_req_cnt=0 http_req_rate(1
Could have thought of that before…
Here’s the valgrind info after installing the debug symbols.
root@haproxy-1:/var/crash# valgrind haproxy -d -f
/vagrant/configs/crasht-test.cfg
==4802== Memcheck, a memory error detector
==4802== Copyright (C) 2002-2013, and GNU GPL'd, by Julian Seward et al.
Hi!
I am seeing a segfault upon the first request coming through the configuration
below.
My intention is to enforce a) a total request limit per minute and b) a
separate limit for certain API paths. For that purpose, in addition to the
be_api_external table, which I intend to use for the tota
Similar issue was fixed in 67dad27, but the fix is incomplete. Crash still
happened when utilizing req.fhdr() and sending exactly MAX_HDR_HISTORY
headers.
This fix needs to be backported to 1.5 and 1.6.
Signed-off-by: Nenad Merdanovic
---
src/proto_http.c | 7 +--
1 file changed, 5 insertio
Hi Willy,
On 2016-03-25 18:17, Willy Tarreau wrote:
On Fri, Mar 25, 2016 at 01:53:50PM +0100, Willy Tarreau wrote:
I think it's even different (but could be wrong) since Christian spoke
about counters suddenly doubling. The issue you faced Sylvain which I
still have no idea how to fix unfortuna
26 matches
Mail list logo