RE: [cross-project-issues-dev] outage of bugs.eclipse.org for Mylyn user

["Mik Kersten" <beatmik@xxxxxxx>]

> -----Original Message-----
> From: cross-project-issues-dev-bounces@xxxxxxxxxxx [mailto:cross-
> project-issues-dev-bounces@xxxxxxxxxxx] On Behalf Of Denis Roy
...
> FWIW, the Mylyn/Bugzilla issue was fixed shortly after this e-mail was
sent.

Yes, we were very happy to see the fix.  Based on bug postings it looks like
the outage was from 10:18 to 20:23 Eastern time (unfortunately we did not
notice it until 14:50).  New users or fresh workspaces trying to connect to
bugs.eclipse.org were the most problematic (see below), which is why I
raised this alarm on cross-projects (under the assumption that Europa Fall
Update triggers fresh downloads). 
 
> Mik Kersten wrote:
> > Some users have already perceived this as a broken Eclipse
> > Fall Update
> If that's the case, perhaps Mylyn needs to be more robust in dealing
> with "normal" conditions that can break Eclipse, such as network
> congestion or Bugzilla being offline for maintenance?  I've opened this
> to track the effort:
> 
> https://bugs.eclipse.org/bugs/show_bug.cgi?id=205295

My detailed response is on that bug.  Mylyn already fails gracefully in all
such scenarios that we have know of, including those listed.  A key problem
here was that new users could not start using the tool because Mylyn cannot
work with Bugzilla installs that block access to their configuration, which
bugs.eclipse.org was temporarily doing.  While most existing users would
have not noticed, since their locally cached configuration was being used,
new users would see the bugs.eclipse.org was working via web browser, but
that Mylyn could not connect, whereas it previously could.  

The problem is now resolved, and I'll provide additional follow up on how I
hope that we can avoid this kind of unnecessary outage on Bjorn's
"eclipse.org server usage patterns" message.

Mik