My ISP changed some configuration but it did not improve the situation.
They asked me if docs.python.org was the only site that caused
troubles and I tried to find others but was not able. So it seems that
it is only docs.python.org.
Frédéric
___
users ma
On 04/13/17 12:53, fedora wrote:
> openssl s_client -connect hostname:portnumber -starttls service
FYI, is you look back on the thread the port number/service being tested
is 443/https.
-starttls prot - use the STARTTLS command before starting TLS
for those protocols that suppo
> for a try with SSL: did you ever use
> openssl s_client -connect hostname:portnumber
this one with docs.python.org showed error 104.
> openssl s_client -connect hostname:portnumber -starttls service
Would it be the same?
Frédéric
___
users mailing l
Hi Frédéric
for a try with SSL: did you ever use
openssl s_client -connect hostname:portnumber
or
openssl s_client -connect hostname:portnumber -starttls service
suomi
On 04/12/2017 10:49 AM, Frédéric Bron wrote:
I contacted my ISP. They do not want to look at the problem if I
cannot show that
I contacted my ISP. They do not want to look at the problem if I
cannot show that I loose packets. But I tried to ping doc.python.org
and never get an error. Apparently, it is related to SSL? What could I
show as a "as simple as possible example"?
They say there are plenty of thing involving IPv6
On Wed, 2017-04-12 at 06:45 +0800, Ed Greshko wrote:
> On 04/12/17 06:10, Patrick O'Callaghan wrote:
> > On Wed, 2017-04-12 at 05:05 +0800, Ed Greshko wrote:
> > > (I think there may be a claim of redundancy to be made
> > > about that paring of words)
> >
> > I think there may be a claim of ortho
On 04/12/17 06:10, Patrick O'Callaghan wrote:
> On Wed, 2017-04-12 at 05:05 +0800, Ed Greshko wrote:
>> (I think there may be a claim of redundancy to be made
>> about that paring of words)
> I think there may be a claim of orthography against that spelling.
> Pedantic, moi?
>
I see the knives are
On Wed, 2017-04-12 at 05:05 +0800, Ed Greshko wrote:
> (I think there may be a claim of redundancy to be made
> about that paring of words)
I think there may be a claim of orthography against that spelling.
Pedantic, moi?
poc
___
users mailing list -- u
On 04/12/17 04:48, Rick Stevens wrote:
> On 04/11/2017 11:42 AM, Gordon Messmer wrote:
>> On 04/10/2017 11:54 PM, Ed Greshko wrote:
You mean my ISP does not handle IPv6 correctly, right?
For the short term, is there any harm to stick to IPv4?
>>> In my opinion, it is not clear that it is
On 04/11/2017 11:42 AM, Gordon Messmer wrote:
> On 04/10/2017 11:54 PM, Ed Greshko wrote:
>>> You mean my ISP does not handle IPv6 correctly, right?
>>> For the short term, is there any harm to stick to IPv4?
>> In my opinion, it is not clear that it is your ISP as there are probably
>> many boxes
On 04/10/2017 11:54 PM, Ed Greshko wrote:
You mean my ISP does not handle IPv6 correctly, right?
For the short term, is there any harm to stick to IPv4?
In my opinion, it is not clear that it is your ISP as there are probably
many boxes between your system and the destination. Some owned by you
On 04/10/2017 11:37 PM, Ed Greshko wrote:
Yes, openssl does take ipv6 addresses. You just need to enclose in
brackets
openssl s_client -connect [2a04:4e42::223]:443
works fine
Huh. Seems Fedora patches openssl for that[1]. Upstream openssl
doesn't support IPv6. Adding to my confusion, the
On 04/11/17 14:37, Frédéric Bron wrote:
>>> In your opinion, is this an issue with fedora or with my ISP (which is
>>> OVH)?
>> Probably your ISP, given the packet sequence you saw. Unless someone else
>> can replicate the problem with that IP address. I might try tomorrow, but
>> openssl's s_cli
>> In your opinion, is this an issue with fedora or with my ISP (which is
>> OVH)?
>
> Probably your ISP, given the packet sequence you saw. Unless someone else
> can replicate the problem with that IP address. I might try tomorrow, but
> openssl's s_client doesn't take ipv6 addresses as argument
On 04/11/17 14:34, Gordon Messmer wrote:
> Probably your ISP, given the packet sequence you saw. Unless someone
> else can replicate the problem with that IP address. I might try
> tomorrow, but openssl's s_client doesn't take ipv6 addresses as
> arguments, so testing is more involved.
Yes, op
On 04/10/2017 11:09 PM, Frédéric Bron wrote:
It would be useful to disable IPv6 for this connection and try again.
So applied the command proposed by Ed as root:
$ echo 1 > /proc/sys/net/ipv6/conf/(name of interface)/disable_ipv6
It might be simpler to disable IPv6 on one connection. From you
> It would be useful to disable IPv6 for this connection and try again.
So applied the command proposed by Ed as root:
$ echo 1 > /proc/sys/net/ipv6/conf/(name of interface)/disable_ipv6
and now I never have any error, the page always shows in firefox,
konqueror or chrome.
Also openssl s_client -
On 04/10/2017 09:30 PM, Ed Greshko wrote:
May I suggest that IP address be used instead of the host name?
I specifically wanted to use the hostname because I assumed that if this
were an IPv6 vs IPv4 sort of problem, we'd see that difference in
successful and unsuccessful connections. We di
On 04/10/2017 08:56 PM, Frédéric Bron wrote:
# tcpdump -nn port 443
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on enp62s0u1u4, link-type EN10MB (Ethernet), capture size 262144 bytes
05:38:34.536393 IP6 2001:41d0:fe0b:8000:9a64:ff80:35b8:7d03.60038 >
2a04
On 04/11/17 00:25, Gordon Messmer wrote:
> On 04/09/2017 11:46 PM, Frédéric Bron wrote:
>> First try and apprently maybe a first error (write:errno=104):
>
>
> OK, errno 104 is ECONNRESET (Connection reset by peer). That might be
> because some firewall or router in between your system and the ser
On 04/11/17 11:56, Frédéric Bron wrote:
> apparently, although I ticked "IPv4 required for this connexion", IPv6 is
> used.
Just a quick note on the meaning of that check box.
It means that in order for that interface to be marked as UP it needs to
have an IPv4 address. That's all it means. I
> Either way, I think the possibility of an IPv6 problem is worth
> investigating. Close all of your applications, then open two terminals. In
> one, run:
>
> $ sudo tcpdump -nn port 443
>
> In the other, run the same s_client command to connect to the python docs
> server. Send us the output of
On 04/09/2017 11:46 PM, Frédéric Bron wrote:
First try and apprently maybe a first error (write:errno=104):
OK, errno 104 is ECONNRESET (Connection reset by peer). That might be
because some firewall or router in between your system and the server is
sending a TCP reset packet to disrupt th
> Just because I am out of ideas at the moment could you try disabling
> IPv6 to see if that helps? You can do it without rebooting by (as root)
>
> echo 1 > /proc/sys/net/ipv6/conf/(name of interface)/disable_ipv6
> And then see if this makes any difference.
in fact, some time ago, I contacted m
> nss doesn't have a test client that I'm aware of, but since you're having
> trouble with an OpenSSL application, you might be able to get more
> information by using the "s_client" app:
>
> openssl s_client -connect docs.python.org:443
First try and apprently maybe a first error (write:errno=104
On 4/7/17 3:39 PM, Ed Greshko wrote:
On 04/07/17 12:32, Frédéric Bron wrote:
gnutls has this on its web page:
Support for TLS 1.2, TLS 1.1, TLS 1.0, and SSL 3.0 protocols
I have
Name: gnutls
Version : 3.5.10
Release : 1.fc25
installed.
Do you have gnutls installed?
gnutls:
Ver
OK, so you're having intermittent problems with both firefox and
konqueror within a fresh user account. That's useful information,
because those two browsers use completely separate SSL implementations.
Mozilla develops their own "nss" library for encryption, and Konqueror
uses OpenSSL.
nss
On 04/07/17 22:44, Frédéric Bron wrote:
>> I think I may have lost part of the thread due to my "expiration"
>> policy. Just to verify. Both your Win and F25 systems are connected
>> via Wifi?
> No wired connexion.
>
OK. If I am not mistaken the slow transfer rates in previous posts
have be
> I think I may have lost part of the thread due to my "expiration"
> policy. Just to verify. Both your Win and F25 systems are connected
> via Wifi?
No wired connexion.
> And, when they connect they are both connected to the same SSID?
same ISP/box.
> Assuming Wifi again, is the brand of HW
On 04/07/17 18:00, Patrick O'Callaghan wrote:
> I'm wondering if this could be related to MTU issues, which in the past
> have sometimes caused very mysterious behaviour. Not sure how to check
> though. Maybe use tracepath (not traceroute), or ping with different
> size packets might show something
On Fri, 2017-04-07 at 17:51 +0800, Ed Greshko wrote:
> On 04/07/17 16:32, Frédéric Bron wrote:
> > > One would think that if this really was a "fedora" issue then many
> > > others would be seeing the same problem. I've not seen evidence for that.
> >
> > I have seen another thread with very slow
On 04/07/17 16:32, Frédéric Bron wrote:
>> One would think that if this really was a "fedora" issue then many
>> others would be seeing the same problem. I've not seen evidence for that.
> I have seen another thread with very slow wifi connexion. I wonder if
> it is not the same issue.
I think I
> One would think that if this really was a "fedora" issue then many
> others would be seeing the same problem. I've not seen evidence for that.
I have seen another thread with very slow wifi connexion. I wonder if
it is not the same issue.
> You said you have a Windows system that doesn't have
On 04/07/17 12:32, Frédéric Bron wrote:
gnutls has this on its web page:
Support for TLS 1.2, TLS 1.1, TLS 1.0, and SSL 3.0 protocols
I have
Name: gnutls
Version : 3.5.10
Release : 1.fc25
installed.
Do you have gnutls installed?
> gn
>> > gnutls has this on its web page:
>> > Support for TLS 1.2, TLS 1.1, TLS 1.0, and SSL 3.0 protocols
>> >
>> > I have
>> > Name: gnutls
>> > Version : 3.5.10
>> > Release : 1.fc25
>> > installed.
>> >
>> > Do you have gnutls installed?
gnutls:
Version : 3.5.10
Release :
On Thu, 6 Apr 2017 14:52:48 -0700
stan wrote:
> On Thu, 6 Apr 2017 14:26:54 -0700
> stan wrote:
>
> > On Thu, 6 Apr 2017 22:29:30 +0200
> > Frédéric Bron wrote:
> >
> > > How do I now if I have TLS 1.2?
> >
> > gnutls has this on its web page:
> > Support for TLS 1.2, TLS 1.1, TLS 1.0
On Thu, 6 Apr 2017 14:26:54 -0700
stan wrote:
> On Thu, 6 Apr 2017 22:29:30 +0200
> Frédéric Bron wrote:
>
> > How do I now if I have TLS 1.2?
>
> gnutls has this on its web page:
> Support for TLS 1.2, TLS 1.1, TLS 1.0, and SSL 3.0 protocols
>
> I have
> Name: gnutls
> Version
On Thu, 6 Apr 2017 22:29:30 +0200
Frédéric Bron wrote:
> How do I now if I have TLS 1.2?
gnutls has this on its web page:
Support for TLS 1.2, TLS 1.1, TLS 1.0, and SSL 3.0 protocols
I have
Name: gnutls
Version : 3.5.10
Release : 1.fc25
installed.
Do you have gnutls installed
> Before you condemn F25, try some other browser than Konqueror and make
> sure you update your system fully.
My system is always up to date. I tested with firefox and konqueror. I
normally use firefox.
> The SSL negotiation may have to do
> with an outdated SSL or TLS library (that site uses TLS
> Can you create an entirely new user on the F25 system and replicate the
> problem there?
>
> What did you copy from opensuse to F25 when you made the switch (if
> anything)?
I installed F25 from scratch: no upgrade from oan lder fedora release.
The account was completely new on one computer (I h
On 04/05/2017 10:49 PM, Frédéric Bron wrote:
Where could it come from?
Can you create an entirely new user on the F25 system and replicate the
problem there?
What did you copy from opensuse to F25 when you made the switch (if
anything)?
___
user
On Thu, 6 Apr 2017 07:49:43 +0200
Frédéric Bron wrote:
> I have now the proof my connexion issues come from fedora.
> Today, again, from firefox, I could not reach this page:
> https://docs.python.org/2/library/shutil.html
> From konqueror, I couldn't reach it but got the same SSL error
> messag
On 04/05/2017 10:49 PM, Frédéric Bron wrote:
>> I often experience very slow internet surf. No idea where it comes
>> from, just that it started when I switched from opensuse to F25.
>>
>> Today, I may have obtained a clue:
>>
>> I normally use firefox.
>> Today I got the following issue:
>> - sear
> I often experience very slow internet surf. No idea where it comes
> from, just that it started when I switched from opensuse to F25.
>
> Today, I may have obtained a clue:
>
> I normally use firefox.
> Today I got the following issue:
> - searched in google for "python os.path"
> - clicked on th
> Nonetheless, I'm on the side of there being some sort of dns issue,
> especially if using any ad blockers (404's, timeouts, etc).
Could you develop a little bit? I have AdBlock Plus extension in firefox.
Frédéric
___
users mailing list -- users@lists.
On Thu, 23 Mar 2017 12:38:07 +1030
Tim wrote:
> Way back in the past, I used to do that kind of thing, but found that
> my PCs struggled with moderately large hosts files, never mind one
> that big. It made all domain name look-ups slow, as it everything
> had to trawl through that hosts file, f
stan:
>> One thing I know slows down browsing is the way sites outsource much
>> of their content. The browser has to access many sites on the web to
>> put together a page for you to view. Blocking this as much as
>> possible not only speeds up page loading, but also hinders tracking
>> sites.
On Wed, 22 Mar 2017 20:36:59 -0400
Peter Skensved wrote:
> > One thing I know slows down browsing is the way sites outsource much
> > of their content. The browser has to access many sites on the web
> > to put together a page for you to view. Blocking this as much as
> > possible not only spee
> One thing I know slows down browsing is the way sites outsource much
> of their content. The browser has to access many sites on the web to
> put together a page for you to view. Blocking this as much as
> possible not only speeds up page loading, but also hinders tracking
> sites.
mvps is
On 03/22/2017 05:05 PM, Rick Stevens wrote:
Standard rule: Do NOT make your page rendering block on external events,
feeds, whatnot. People will blame your page, not the events that are
blocking it.
Especially if they've blocked those sites.
___
users
On 03/22/2017 04:36 PM, Tim wrote:
> Allegedly, on or about 22 March 2017, stan sent:
>> One thing I know slows down browsing is the way sites outsource much
>> of their content. The browser has to access many sites on the web to
>> put together a page for you to view. Blocking this as much as
>>
Allegedly, on or about 22 March 2017, stan sent:
> One thing I know slows down browsing is the way sites outsource much
> of their content. The browser has to access many sites on the web to
> put together a page for you to view. Blocking this as much as
> possible not only speeds up page loading
On Wed, 22 Mar 2017 13:12:19 -0700
Joe Zeff wrote:
> Why not simply run it in a terminal? It doesn't *have* to be
> maximized, you know.
True. Just a habit that I avoid having root running in X for long
periods of time.
___
users mailing list -- user
On 03/22/2017 12:12 PM, stan wrote:
On Wed, 22 Mar 2017 12:05:43 -0700
stan wrote:
One thing I know slows down browsing is the way sites outsource much
of their content. The browser has to access many sites on the web to
put together a page for you to view. Blocking this as much as
possible
On 03/22/2017 12:05 PM, stan wrote:
There are probably better ways to do this, but you could install iftop,
open a virtual console, login as root, and leave iftop running. When
the problem occurs, flip to the virtual console, and see a snapshot of
the current connections. That might point you t
On Wed, 22 Mar 2017 12:05:43 -0700
stan wrote:
> One thing I know slows down browsing is the way sites outsource much
> of their content. The browser has to access many sites on the web to
> put together a page for you to view. Blocking this as much as
> possible not only speeds up page loading
On Wed, 22 Mar 2017 19:10:21 +0100
Frédéric Bron wrote:
> > Given your description of what you've already tried to isolate the
> > problem, I think it is something in your web access, either DNS or
> > ISP.
>
> I also suspect that but when I run their speed test everything is
> fine!
There ar
> It could be the way Firefox is handling SSL connections
> I'd suggest ensuring you are running the latest Firefox
I update F25 very often, thanks.
Frédéric
___
users mailing list -- users@lists.fedoraproject.org
To unsubscribe send an email to users-l
> Given your description of what you've already tried to isolate the
> problem, I think it is something in your web access, either DNS or
> ISP.
I also suspect that but when I run their speed test everything is fine!
Frédéric
___
users mailing list -- u
On 03/22/2017 09:59 AM, stan wrote:
> On Wed, 22 Mar 2017 17:07:49 +0100
> Frédéric Bron wrote:
>
>> I often experience very slow internet surf. No idea where it comes
>> from, just that it started when I switched from opensuse to F25.
>>
>> Today, I may have obtained a clue:
>>
>> I normally use
On Wed, 22 Mar 2017 17:07:49 +0100
Frédéric Bron wrote:
> I often experience very slow internet surf. No idea where it comes
> from, just that it started when I switched from opensuse to F25.
>
> Today, I may have obtained a clue:
>
> I normally use firefox.
> Today I got the following issue:
>
I often experience very slow internet surf. No idea where it comes
from, just that it started when I switched from opensuse to F25.
Today, I may have obtained a clue:
I normally use firefox.
Today I got the following issue:
- searched in google for "python os.path"
- clicked on the first linked h
62 matches
Mail list logo
