Hi Christopher,
On Mon, Feb 19, 2018 at 03:24:09PM +0100, Christopher Faulet wrote:
> Someone on discourse reports a problem with this patch:
>
> https://discourse.haproxy.org/t/random-sa-errors-with-haproxy-1-8-3/2116/6
>
> I asked him to test the attached patch. But It could be cool to have mo
Le 14/02/2018 à 18:53, Willy Tarreau a écrit :
On Wed, Feb 14, 2018 at 06:20:42PM +0100, Mateusz Malek wrote:
Hi,
On 14.02.2018 17:53, Willy Tarreau wrote:
On Wed, Feb 14, 2018 at 05:29:57PM +0100, Olivier Houchard wrote:
What about what's attached, instead ?
I think it should work. Mateusz,
On Wed, Feb 14, 2018 at 06:20:42PM +0100, Mateusz Malek wrote:
> Hi,
>
> On 14.02.2018 17:53, Willy Tarreau wrote:
> > On Wed, Feb 14, 2018 at 05:29:57PM +0100, Olivier Houchard wrote:
> > > What about what's attached, instead ?
> > I think it should work. Mateusz, care to give it a try to confirm
Hi,
On 14.02.2018 17:53, Willy Tarreau wrote:
On Wed, Feb 14, 2018 at 05:29:57PM +0100, Olivier Houchard wrote:
What about what's attached, instead ?
I think it should work. Mateusz, care to give it a try to confirm ?
If OK, I'll merge it.
I confirm, with this patch applied problem is gone.
On Wed, Feb 14, 2018 at 05:29:57PM +0100, Olivier Houchard wrote:
> Hi Willy,
>
> On Tue, Feb 13, 2018 at 08:05:44PM +0100, Willy Tarreau wrote:
> > Hi Olivier,
> > Such type of construct tends to scare me (probably because I'm not reading
> > the whole code). It means we're supposed to set an err
Hi Willy,
On Tue, Feb 13, 2018 at 08:05:44PM +0100, Willy Tarreau wrote:
> Hi Olivier,
> Such type of construct tends to scare me (probably because I'm not reading
> the whole code). It means we're supposed to set an error by default unless
> we pass by a specific path. I fear that we'll get futur
Hi Olivier,
On Tue, Feb 13, 2018 at 06:07:36PM +0100, Olivier Houchard wrote:
> Hi Emmanuel,
>
> On Tue, Feb 13, 2018 at 05:40:00PM +0100, Emmanuel Hocdet wrote:
> > Hi Olivier
> >
> > > Le 13 févr. 2018 à 15:27, Olivier Houchard a
> > > écrit :
> > >
> > > Thanks a lot for the detailed analy
Hi Emmanuel,
On Tue, Feb 13, 2018 at 05:40:00PM +0100, Emmanuel Hocdet wrote:
> Hi Olivier
>
> > Le 13 févr. 2018 à 15:27, Olivier Houchard a écrit :
> >
> > Thanks a lot for the detailed analyze, and sorry for the late answer.
> > You're probably right, SSL_ERROR_SYSCALL shouldn't be treated a
Hi Olivier
> Le 13 févr. 2018 à 15:27, Olivier Houchard a écrit :
>
> Thanks a lot for the detailed analyze, and sorry for the late answer.
> You're probably right, SSL_ERROR_SYSCALL shouldn't be treated as an
> unrecoverable error.
> So, what you basically did was something equivalent to the pa
On Tue, Feb 13, 2018 at 05:29:21PM +0100, Mateusz Malek wrote:
> Hi,
>
> On 13.02.2018 15:27, Olivier Houchard wrote:
> > Thanks a lot for the detailed analyze, and sorry for the late answer.
> > You're probably right, SSL_ERROR_SYSCALL shouldn't be treated as an
> > unrecoverable error.
> > So, w
Hi,
On 13.02.2018 15:27, Olivier Houchard wrote:
> Thanks a lot for the detailed analyze, and sorry for the late answer.
> You're probably right, SSL_ERROR_SYSCALL shouldn't be treated as an
> unrecoverable error.
> So, what you basically did was something equivalent to the patch attached ?
Yeah,
Hi guys,
On Sat, Feb 10, 2018 at 06:26:42PM +0100, Mateusz Małek wrote:
> Hi everyone,
>
> I've narrowed down my problem down to the same commit as Tomek Gacek -
> c2aae74f010f97a3415542fe649198a5d3be1ea8 (MEDIUM: ssl: Handle early data
> with OpenSSL 1.1.1), so I guess it may be related. In my c
Hi Mateusz,
I'm CCing Emeric (SSL maintainer) and Olivier (who added early-data support),
and responding to some points below.
On Sat, Feb 10, 2018 at 06:26:42PM +0100, Mateusz Malek wrote:
> Hi everyone,
>
> I've narrowed down my problem down to the same commit as Tomek Gacek -
> c2aae74f010f97
Hi everyone,
I've narrowed down my problem down to the same commit as Tomek Gacek -
c2aae74f010f97a3415542fe649198a5d3be1ea8 (MEDIUM: ssl: Handle early data
with OpenSSL 1.1.1), so I guess it may be related. In my case, since
upgrade to 1.8, some responses from some backends (not sure what exa
Hi Willy
On 2018-02-03 10:05, Willy Tarreau wrote:
Hi Tomek,
On Sat, Feb 03, 2018 at 08:47:35AM +0100, Tomek Gacek wrote:
I have same issue. It's pretty random as I would say about 60-70% requests
are OK, but rest is failing. I compiled all 1.8 versions and was able to
isolate this a little bi
Hi Tomek,
On Sat, Feb 03, 2018 at 08:47:35AM +0100, Tomek Gacek wrote:
> I have same issue. It's pretty random as I would say about 60-70% requests
> are OK, but rest is failing. I compiled all 1.8 versions and was able to
> isolate this a little bit. It's fine up to 1.8.0-dev3 branch and it's
> f
Hi
On 2018-01-17 11:37, Bart Geesink wrote:
Hi,
On 01/17/2018 10:16 AM, Christopher Faulet wrote:
Le 16/01/2018 à 16:18, Lukas Tribus a écrit :
Hello Christopher,
On 16 January 2018 at 15:01, Bart Geesink
wrote:
Hi,
We have an issue in haproxy > 1.8 on CentOS when using SSL in the server
Hi,
On 01/17/2018 10:16 AM, Christopher Faulet wrote:
> Le 16/01/2018 à 16:18, Lukas Tribus a écrit :
>> Hello Christopher,
>>
>>
>> On 16 January 2018 at 15:01, Bart Geesink
>> wrote:
>>> Hi,
>>>
>>> We have an issue in haproxy > 1.8 on CentOS when using SSL in the server
>>> configuration. Hapr
Le 16/01/2018 à 16:18, Lukas Tribus a écrit :
Hello Christopher,
On 16 January 2018 at 15:01, Bart Geesink wrote:
Hi,
We have an issue in haproxy > 1.8 on CentOS when using SSL in the server
configuration. Haproxy sometimes logs a http status code "-1" followed
by the termination_state SDxx.
Hello Christopher,
On 16 January 2018 at 15:01, Bart Geesink wrote:
> Hi,
>
> We have an issue in haproxy > 1.8 on CentOS when using SSL in the server
> configuration. Haproxy sometimes logs a http status code "-1" followed
> by the termination_state SDxx. This happens every few requests. When
>
Hi,
We have an issue in haproxy > 1.8 on CentOS when using SSL in the server
configuration. Haproxy sometimes logs a http status code "-1" followed
by the termination_state SDxx. This happens every few requests. When
using one backend, the clients don't notice it. When using multiple
backends, thi
21 matches
Mail list logo
