[Teammetrics-discuss] Updated data

[Andreas Tille]

Hi Sukhbir,

On Sat, Aug 03, 2013 at 12:32:00PM -0400, Sukhbir Singh wrote:
> 
> Multiple things seem to have broken now. For example, see this from
> `liststat.log':
> 
> 2013-08-03 15:47:38,083 ERROR: Unable to fetch mailing list archive <urlopen error [Errno -2] Name or service not known>
> 2013-08-03 15:47:38,083 INFO: 	List 17 of 18
> 2013-08-03 15:47:38,083 INFO: Reading http://lists.alioth.debian.org/pipermail/pkg-kolab-devel
> 2013-08-03 15:48:34,128 ERROR: Unable to fetch mailing list archive <urlopen error [Errno -2] Name or service not known>
> 2013-08-03 15:48:34,128 INFO: 	List 18 of 18
> 2013-08-03 15:48:34,128 INFO: Reading http://lists.alioth.debian.org/pipermail/pkg-horde-hackers
> 2013-08-03 15:49:30,170 ERROR: Unable to fetch mailing list archive <urlopen error [Errno -2] Name or service not known>
> 
> So given the above log, I have two theories:
> 
> - We have a connection issue from the VPS, because the above URLs are
>   working fine otherwise.

It is the stupid issue with broken DNS.  I just reported it previously
to you but have not yet been able to solve this.  I'll take care when
beeing at debconf.  For the moment I added lists.alioth.debian.org to
/etc/hosts to work around this for the moment.

> On that related note, I have seen that it
>   takes quite some time for me to log in to the VPS. Can these issues be
>   related?

Same observation from my side and yes, I'm afraid that it is also
connected to the DNS issue.
 
> - Also equally probable: I think my cronjob ran at the same time as you
>   started the script. So that caused the problem because of the same
>   script running twice.

Hmmm, I also have wondered what might happen if the script might run
twice at the same time.  I had a similar problem in the Blends
websentinel and solved it this way:

   http://anonscm.debian.org/gitweb/?p=blends/website.git;a=blob;f=webtools/blendstasktools.py
   look for LockBlendsTools()

I think we should implement something like this (or some more clever
mechanism).

 
> Both of the above are equally probable and I will check more based on
> your reply.

I guess it should be possible to try to reimport the data now.  I'm
afraid it is wrong since April or Mai when I observed the DNS problem
the first time.  To stupid that I have no idea. :-(

Kind regards

        Andreas.

-- 
http://fam-tille.de