Operational Meeting

Apologies were received from Ken, Craig and Simon.

Toby suggested that the section on LDAP logging should say that the default logging level is as low as it'll go while still logging anything at all, but that it can still cause problems on machines which do lots of lookups.

The following item was removed from the actions list:

• Look at LDAP logging

Report from Computing Executive Group

Frontline Support Issues

• A new LaTeX person is in the pipeline. Alison will think about how we could cooperate with other Schools.

Reports from units.

• Infrastructure.

  4. JCMB 2501

     Neil will phone E&B again to see what, if anything, is happening about the air-conditioning unit.

  5. OpenLDAP proxy-caching

     Toby and Iain will discuss condor and the clusters. There may also be a few other odd machines where it would be useful to test the proxy-caching.

• Managed Platform.

  1. New PXE service

     Stephen would like to have a "non-install" afternoon some time soon so that he can switch over an entire wire to the new PXE mechanism for testing. He'll suggest one. In due course, the location of the PXE servers will be reviewed -- they're currently on the inf-unit's network infrastructure machines -- and the way they're configured will be changing slightly.

• Research and Teaching.

  1. Examlockdown script broke

     inf-unit will add an rfe replacement to its project wish-list. (It's 107.)

  • There's now a way to have people opted-out of all database-generated mailing lists; previously it could be done, somewhat tediously, by removing them from lists individually.

• Services.

  3. inf-general mailing list and newsgroup

     The newsgroup now exists. We think it should be bidirectionally gatewayed to the mailing list. Membership should be for the School office to decide. It was noted that the eduni groups are visible throughout the University, but not outside.

• User Support.

Topics for discussion

• Research support

  Tim drew the meeting's attention to the procedure for obtaining research support (linked from the right-side menu bar of the "systems" page). In summary, it indicates how researchers can obtain additional effort to implement things which aren't part of the commodity base. Anyone handling RT tickets from researchers is asked to bear this procedure in mind, and to query the rat unit as necessary.

• Institute liaison

  As originally devised, institute liaison COs had two roles:

  1. to provide a problem-escalation mechanism; and
  2. to be proactive in representing institutes' needs to the CO body.

  The first of these is no longer appropriate. The question is, whether and in what form the second should be continued? The general feeling of the meeting was that while we don't want to confuse the lines of responsibility, we do feel more isolated in the new building and that in itself was a good reason to have some kind of liaison in place, even if it might have to be different for different institutes.

• eDiary

  There was some discussion on whether and how we should be promoting the use of eDiary within the School. Particular questions were: off-line working and how to synchronise with other diaries and organisers; how to mix personal and work-related entries; and how to express preferences regarding open slots. IS have packaged up a client for SL5, and we should at least make it available to try out.

• DNS round-robin issues

  inf-unit had come across a well-known problem regarding round-robin DNS entries being unexpectedly sorted by clients. This had turned out to be a glibc attempt to implement RFC 3484. It was concluded that this is a problem, and that the inf-unit should look at ways to avoid it (by, for example, /etc/gai.conf changes).

• It was decided to defer the discussion on the School Computing Strategy again.

AOCB

The next meeting will be on Wednesday the 10th of December at 10:00, apparently in IF-5.42, chaired by Tim Colles.