Bug#741104: [Popcon-developers] Bug#741104: popularity-contest: Silently fails to send anything if no MTA configured

2016-03-20 Thread Francesco Poli
On Sun, 20 Mar 2016 15:10:41 +0100 Bill Allombert wrote:

> On Sun, Mar 20, 2016 at 11:49:34AM +0100, Francesco Poli wrote:
> > On Mon, 13 Oct 2014 22:10:20 +0200 Bill Allombert wrote:
> > 
> > Hello Bill,
> > I use popularity-contest on most Debian boxes I administer and I have
> > experienced this issue where the MTA is configured to only deal with
> > local mail.
> 
> Please do not hijack other people bug report...

Sorry, it seemed to me that my issue was the same as (or, at least,
closely related to) the one discussed in this bug report.
I apologize, if my mistake caused any trouble.

> 
> > On one such system, I get the bounce basically once a week.
> > Since I haven't changed the default configuration of
> > popularity-contest, this means that the HTTP method almost always fails
> > and the SMTP method is tried (in vain). Do I understand correctly?
> 
> The problem addressed in this bug report is that popcon is assuming that if
> /usr/sbin/sendmail report success, then the email will be delivered, so
> it marks the submission as successful and do not resend it again.
> 
> > Please note that the system is not up and running 24/7, and uses
> > anacron: hence, the HTTP method is not tried always at the same time.
> > Nonetheless, it seems to almost always fail!   :-(
> 
> Do you know why ?

I suspect that this is due to somewhat unreliable DNS servers.
I noticed that sometimes I have to query the DNS servers more than
once, before I get a meaningful response.
Maybe the HTTP method fails, because it cannot resolve
popcon.debian.org ...

> Is the network available already when anacron run ?

Yes, it is, but probably with flaky DNS servers...

> If you cannot report neither by HTTP nor by SMTP, then there is little
> way for popcon to work.

I don't know, maybe it could retry the HTTP method a second time
(possibly after waiting some 5 seconds)?

> 
> > Is there any progress on the fix you are working on?
> 
> This is a different bug report: #709892
> The server was changed to be able to handle load more efficiently and the
> client part (in testing/unstable) use a randomized report time to spread
> the load.

Sorry, I hadn't found bug #709892, as it is archived.
I assumed that I could not be hit by long-fixed bugs...

> 
> > Questions:
> >  • what would happen, if I added
> >MAILTO=
> >to /etc/popularity-contest.conf?
> 
> Then SMTP will not be attempted and the submission will be marked as
> failed and will be retried the next day.

Thanks for the explanation, maybe it's worth a try.

> 
> Thanks for using popularity-contest!

Thanks to you for maintaining it!
Bye.


-- 
 http://www.inventati.org/frx/
 There's not a second to spare! To the laboratory!
. Francesco Poli .
 GnuPG key fpr == CA01 1147 9CD2 EFDF FB82  3925 3E1C 27E1 1F69 BFFE


pgpmVXTkdGOKK.pgp
Description: PGP signature


Bug#741104: [Popcon-developers] Bug#741104: popularity-contest: Silently fails to send anything if no MTA configured

2016-03-20 Thread Bill Allombert
On Sun, Mar 20, 2016 at 11:49:34AM +0100, Francesco Poli wrote:
> On Mon, 13 Oct 2014 22:10:20 +0200 Bill Allombert wrote:
> 
> Hello Bill,
> I use popularity-contest on most Debian boxes I administer and I have
> experienced this issue where the MTA is configured to only deal with
> local mail.

Please do not hijack other people bug report...

> On one such system, I get the bounce basically once a week.
> Since I haven't changed the default configuration of
> popularity-contest, this means that the HTTP method almost always fails
> and the SMTP method is tried (in vain). Do I understand correctly?

The problem addressed in this bug report is that popcon is assuming that if
/usr/sbin/sendmail report success, then the email will be delivered, so
it marks the submission as successful and do not resend it again.

> Please note that the system is not up and running 24/7, and uses
> anacron: hence, the HTTP method is not tried always at the same time.
> Nonetheless, it seems to almost always fail!   :-(

Do you know why ? Is the network available already when anacron run ?
If you cannot report neither by HTTP nor by SMTP, then there is little
way for popcon to work.

> Is there any progress on the fix you are working on?

This is a different bug report: #709892
The server was changed to be able to handle load more efficiently and the
client part (in testing/unstable) use a randomized report time to spread
the load.

> Questions:
>  • what would happen, if I added
>MAILTO=
>to /etc/popularity-contest.conf?

Then SMTP will not be attempted and the submission will be marked as
failed and will be retried the next day.

Thanks for using popularity-contest!
-- 
Bill. 

Imagine a large red swirl here. 



Bug#741104: [Popcon-developers] Bug#741104: popularity-contest: Silently fails to send anything if no MTA configured

2014-03-09 Thread Bill Allombert
On Sun, Mar 09, 2014 at 07:28:14AM +0800, Paul Wise wrote:
 On Sun, Mar 9, 2014 at 1:42 AM, Bill Allombert wrote:
 
  popularity-contest is able to detect whether a MTA is installed or not, but 
  not whether
  the MTA is able to send email to the popcon.debian.org server.
 
 If you want to change that, one possibility is to use VERP (with
 libmail-verp-perl) to read bounce messages and detect that mail
 sending should be disabled.

I doubt it. How popularity-contest would be supposed to receive the bounce in 
the first
place ?

Cheers,
-- 
Bill. ballo...@debian.org

Imagine a large red swirl here. 


-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org



Bug#741104: [Popcon-developers] Bug#741104: popularity-contest: Silently fails to send anything if no MTA configured

2014-03-08 Thread Paul Wise
On Sun, Mar 9, 2014 at 1:42 AM, Bill Allombert wrote:

 popularity-contest is able to detect whether a MTA is installed or not, but 
 not whether
 the MTA is able to send email to the popcon.debian.org server.

If you want to change that, one possibility is to use VERP (with
libmail-verp-perl) to read bounce messages and detect that mail
sending should be disabled.

-- 
bye,
pabs

http://wiki.debian.org/PaulWise


-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org