[Gt-eos] Fwd: New globus-gssapi-gsi version default to TLSv1.2

Maarten Litmaath Maarten.Litmaath at cern.ch
Mon Sep 24 19:50:16 CEST 2018


Hi all,
the FTS pilot at CERN is a production service that runs with EPEL-testing enabled
and it actually caught the problem, for which a ticket was then opened ~2 weeks ago.

Unfortunately, the people first involved with that ticket could not really be expected to
recognize the wider scope of the problem: after all, just a single SE did not work and
nobody screamed...  By the time the FTS team finally got involved, the update had
just gone to production.  BTW, Friday may not be the best time for updates...

We can look further into this business offline to see if we have the most important
workflows covered with EPEL-testing somewhere.

________________________________________
From: Balazs Konya [balazs.konya at hep.lu.se]
Sent: 24 September 2018 14:28
To: Maarten Litmaath; bbockelm at cse.unl.edu; Mattias Ellert
Cc: Andrea Manzi; End of Support of Globus Toolkit; wlcg-middleware-officer (Group of people with WLCG MW Officer function); Oliver Keeble; operations at egi.eu
Subject: Re: [Gt-eos] Fwd: New globus-gssapi-gsi version default to TLSv1.2

hi Maarten ,all

On 2018.09.21. 18:23, Maarten Litmaath wrote:
> The change actually was announced in the EGI URT meeting of Sep 10:
>
>      https://wiki.egi.eu/wiki/URT:Agenda-2018-09-10#Globus
>
> But nobody who was present realized the potential implications for WLCG...
>
> In the future we need to make sure that such _major_ changes are announced
> in all relevant forums, including the WLCG Operations Coordination meeting.

the announcement should contain a bit more info as well. i read the UMD meeting
notes but did not dig into what exactly got changed with the globus-gssapi-gsi
package

in general, i see a problem how these new packages going through the
EPEL-testing phase without actually being tested/deployed at all.


Balazs



More information about the Gt-eos mailing list