White dot for spacing only
The Dice Project


Operational Meeting

Minutes of the meeting held on Wednesday 26th March 2014 in IF-4.31

Present:
Lindsey Brown, Neil Brown, Roger Burroughes, Alison Downie, Ian Durkacz, Graham Dutton, Jennifer Oxley, Stephen Quinney, Iain Rae, Gordon Reid, George Ross (convener and minutes)
  1. Minutes of the last meeting

    These were accepted.

  2. Actions and blog articles

    Actions discussed:

    Actions added:

    Blog articles suggested:

  3. Report from Computing Executive Group

  4. Reports from units

  5. Topics for discussion

    1. Lessons learned from the 25th March power-down

      Much discussion, with the following points made inter alia:

      • There was considerable disquiet that we hadn't been consulted by E&B before the work was done. Suppose an exam had been affected? As it was, some coursework may require an extension.
      • "Criticality" needs another look. There's too much lumped together: availability, integrity, confidentiality, stopping, starting, migrating, ...
      • Who is supposed to bring up VMs? Who it is is perhaps less important than having a consistent approach so we know what to expect. Should autostart be more used? This all needs to be documented. What happens if machines fail to start?
      • The ongoing "tlsdir" tests caused problems, as many CO desktops were unusable until the LDAP servers were started. It was unfortunate that all the test servers were in the Forum and so affected. Does this suggest that we don't actually want to run a "conventional" LDAP service, as our current slapd-on-all-DICE-machines approach is more robust? Could sssd caches have helped? Ian to look into all this.
      • The server room desktops should all be "full CO" desktops, with all the usual packages and on the <stable> release. They should never be part of any testing arrangement. User-support to arrange.
      • There are merits in having both a <develop> and a <stable> desktop, either real or using virtualbox.
      • Mail got bunged up, which meant that nagios alerts weren't useful and we couldn't easily communicate with users. Neil will consider ways around this problem.
      • IS are going to have a supported jabber service soon. When it comes along we should consider how best to use it.
      • Someone should write a blog article.
      • See also the notes from last time.
  6. Items for the Computing Systems blog

    Forum server room power glitch (see discussion above).

  7. AOCB

  8. Next meeting

    The next meeting will be on 12th April 2014 at 10:00 in IF-4.31, chaired by Craig.


 : Operational : Meetings 

Mini Informatics Logo - Link to Main Informatics Page
Please contact us with any comments or corrections.
Unless explicitly stated otherwise, all material is copyright The University of Edinburgh
Spacing Line