nntp gateway refuses connections

robin_listas · October 29, 2014, 4:23am

Local time, UTC is 1 h less.


<7.4> 2014-10-29 00:33:14 Telcontar fetchnews 26361 - -  warning:  nntp.opensuse.org: connection to 130.57.2.16 failed: Connection refused
<7.4>  2014-10-29 00:33:15 Telcontar fetchnews 26361 - -  warning:  nntp.novell.com: connection to 130.57.2.15 failed: Connection refused
<7.4> 2014-10-29 00:33:21 Telcontar fetchnews 26361 - -  WARNING: some servers have not been queried!
<7.4>  2014-10-29 00:38:06 Telcontar fetchnews 26490 - -  warning:  nntp.opensuse.org: connection to 130.57.2.16 failed: Connection refused
<7.4>  2014-10-29 00:38:07 Telcontar fetchnews 26490 - -  warning:  nntp.novell.com: connection to 130.57.2.15 failed: Connection refused
<7.4> 2014-10-29 00:38:07 Telcontar fetchnews 26490 - -  WARNING: some servers have not been queried!

....

<7.4>  2014-10-29 04:13:05 Telcontar fetchnews 1766 - -  warning:  nntp.opensuse.org: connection to 130.57.2.16 failed: Connection refused
<7.4>  2014-10-29 04:13:05 Telcontar fetchnews 1766 - -  warning:  nntp.novell.com: connection to 130.57.2.15 failed: Connection refused
<7.4> 2014-10-29 04:13:05 Telcontar fetchnews 1766 - -  WARNING: some servers have not been queried!
<7.4>  2014-10-29 04:18:05 Telcontar fetchnews 2048 - -  warning:  nntp.opensuse.org: connection to 130.57.2.16 failed: Connection refused
<7.4>  2014-10-29 04:18:05 Telcontar fetchnews 2048 - -  warning:  nntp.novell.com: connection to 130.57.2.15 failed: Connection refused
<7.4> 2014-10-29 04:18:05 Telcontar fetchnews 2048 - -  WARNING: some servers have not been queried!

malcolmlewis · October 29, 2014, 5:01am

Hi
Known issue, no eta for a fix.

robin_listas · October 29, 2014, 9:56am

Still not working, 08:48 UTC.

I also reported (yesterday) to admin@opensuse.org, got automated response with tick number #4424

malcolmlewis · October 29, 2014, 10:45am

Hi
As explained that’s what no eta means.

deano_ferrari · October 29, 2014, 9:19pm

NNTP is a secondary interface and works as we can. If
you want to ensure you see every message, use the web interface. It’s
clearly written in our FAQ:
https://forums.opensuse.org/faq.php?faq=novfor#faq_nntp

robin_listas · October 30, 2014, 7:18am

On 2014-10-29 21:26, deano ferrari wrote:
>
> robin_listas;2671704 Wrote:
>> Still not working, 08:48 UTC.
>>
>> I also reported (yesterday) to admin@opensuse.org, got automated
>> response with tick number #4424
>
> NNTP is a secondary interface and works as we can. If
> you want to ensure you see every message, use the web interface. It’s
> clearly written in our FAQ:
> https://forums.opensuse.org/faq.php?faq=novfor#faq_nntp

To me, it is primary. If it doesn’t work, I can’t contribute.
Plain simple as that.

I understand that to you it is not. I do not /demand/ more resources.

I would simply appreciate more consideration on part of the staff for
nntp. Like saying in a banner “sorry, we have a problem with nntp, stay
tuned. It is complicated, it will take long”. It would be appreciated a
comment about the type of problem the nntp forum faces, how long it will
take (hours/days/weeks, not a precise figure). And when the issue is
solved, a message about “expect missing messages, or do not expect
them”. The banner could simply point to a news (not nntp) web entry on
the opensuse news site for info and comments, instead of a big banner.

But this is every time forgotten, we have to ask for information every
time, and then you think we are complaining when we are not :-/

Please.

–
Cheers / Saludos,

Carlos E. R.
(from 13.1 x86_64 “Bottle” at Telcontar)

deano_ferrari · October 30, 2014, 7:39am

From an official forum perspective, it is a secondary interface. Arguing about that here is pointless.

I would simply appreciate more consideration on part of the staff for
nntp. Like saying in a banner “sorry, we have a problem with nntp, stay
tuned. It is complicated, it will take long”. It would be appreciated a
comment about the type of problem the nntp forum faces, how long it will
take (hours/days/weeks, not a precise figure). And when the issue is
solved, a message about “expect missing messages, or do not expect
them”. The banner could simply point to a news (not nntp) web entry on
the opensuse news site for info and comments, instead of a big banner.

There was a notice given (forum home page). The fact that you did not check it is not our problem. We’re not always in a position to give times for recoveries from outages.

But this is every time forgotten, we have to ask for information every
time, and then you think we are complaining when we are not :-/

Wrong for the reason given above.

hendersj · October 30, 2014, 11:09pm

On Thu, 30 Oct 2014 06:18:08 +0000, Carlos E. R. wrote:

>> NNTP is a secondary interface and works as we can. If you want to
>> ensure you see every message, use the web interface. It’s clearly
>> written in our FAQ:
>> https://forums.opensuse.org/faq.php?faq=novfor#faq_nntp
>
> To me, it is primary. If it doesn’t work, I can’t contribute.
> Plain simple as that.

No, the plain and simple answer is that it’s a secondary interface, and
so an outage isn’t considered mission critical. Whether it’s your
primary interface or not isn’t what was described, nor is it what is
particularly germane to the discussion.

As we’ve told you time and time again.

On the rare instances that we have an outage (it’s been quite a while
since the last one, actually), we do what we can when we can.

> I understand that to you it is not. I do not /demand/ more resources.

Yet you do - every time we have an outage, we get a complaint from you
that we didn’t post a notice fast enough for your taste. Or that we
didn’t notify you that messages may have been lost.

Yeah, it’s inconvenient. It’s not the end of the world, Carlos.

> I would simply appreciate more consideration on part of the staff for
> nntp. Like saying in a banner “sorry, we have a problem with nntp, stay
> tuned. It is complicated, it will take long”.

Start with the assumption that if you can’t connect and there’s a banner
saying “hey, we have a problem” (which we DID have this time on the web
interface, as soon as it became apparent that there was a serious problem
that was going to take some time to fix).

> It would be appreciated a
> comment about the type of problem the nntp forum faces, how long it will
> take (hours/days/weeks, not a precise figure). And when the issue is
> solved, a message about “expect missing messages, or do not expect
> them”.

Sorry, that’s probably not going to happen. The forum staff here (the
admins and global moderators) don’t have access to the backend servers,
again, AS YOU KNOW - so it gets fixed when the people with access to the
backend servers get a chance to fix it.

Which is what we told you the last time we had a minor NNTP outage.

Less than 1% of our users use NNTP. Yet when there’s an outage, we spend
more and more time assuring a small fraction of those users (current
count is approximately 1).

> The banner could simply point to a news (not nntp) web entry on the
> opensuse news site for info and comments, instead of a big banner.

Again, not likely to happen. By the time we got the notice posted, the
problem would be fixed. I don’t have access to update the banner on
news.o.o, and I don’t think any other member of the forum staff does.

We do have access to post an announcement on the web interface for the
forums, which we did this time around. So maybe look there next time,
and consider that maybe we haven’t noticed that it’s not available (due
to sleeping or other life events) and that’s why a notice hasn’t been
posted yet, instead of complaining 7.5 seconds after the outage starts
that we didn’t post a banner on the web interface.

> But this is every time forgotten, we have to ask for information every
> time, and then you think we are complaining when we are not :-/
>
> Please.

Here is what you should assume:

If you can connect and post, it’s up.
If you cannot connect or post, it’s down.
The chances are that if it’s down, one of two situations is present:

a. We know about it and someone’s working on it. It’ll be fixed when
it can be fixed, so the ETA is “when we figure out the problem and fix
it, it’ll be fixed.”

b. We don’t know about it and will either soon discover it because
members of staff both inside and outside SUSE use NNTP regularly to
connect to these servers. So if the server’s not responding, we’ll find
out as soon as we try to connect.

When it comes up, there’s always a chance that the gateway (ie, the
piece of software that moves the messages back and forth between the NNTP
server and the web forums database - that’s the gateway - the NNTP server
isn’t a gateway itself, it’s an NNTP server) will lose messages. We try
to avoid that. Sometimes we don’t succeed. Sometimes we do. Just
depends on the problem.

This will apply to any future NNTP outage we have. That’s everything you
need to know, and it’s everything that we can tell you about any NNTP
outage. If it’s out, it’ll get fixed as soon as possible. Until then,
it’s offline.

If you want to tell us “hey, it looks like it’s down” on the off-chance a
member of staff hasn’t tried to connect to it, post here from the web
interface. That’s fine. We don’t need any further discussion about it
than that.

Jim

–
Jim Henderson
openSUSE Forums Administrator
Forum Use Terms & Conditions at http://tinyurl.com/openSUSE-T-C