Lindsey, Neil, Toby
These were accepted.
Report from Computing Executive Group
Reports from units.
Alastair has asked E&B whether an evening shutdown for the HV-work is feasible.
Stephen pointed out that manually configured interfaces won't be removed by the updated component.
We are also carrying the ISOs for the various distributions. Perhaps we should let the users know? The package mirrors are also browsable via the web.
Stephen has posted some notes about porting components to F12:
We want to move away from the DICE CVS repository and put everything into subversion. The best way to move your components (and anything else in CVS) is to ask MPU to run the cvs-to-svn import script, this will preserve all history on the files. The LCFG build tools (i.e. lcfg-reltool) do not currently work on F12, getting this fixed is fairly high on my todo list. This is not a showstopper, MPU has successfully ported lots of components by running the build tools on SL5 and then rebuilding the SRPMs on F12 afterwards. Running any svn commands on F12 will instantly upgrade your working copy and you will not be able to use it on SL5 anymore. The only solution would be to delete the copy and do a new checkout on SL5. There is a new 'devel' bucket for F12. This can be used for any development work you like. We would like to keep the standard package buckets a bit cleaner (and smaller) so this is particularly intended for the situation where all you want is to do test installs of various versions of packages via updaterpms. Once you have a "good" package it should then be submitted to the correct bucket as normal. Packages in this bucket are only available to machines following the "develop" release. During the F12 porting period we will be regularly sweeping the devel bucket and removing packages which have newer versions in other buckets. Once F12 has gone "stable" we will be sweeping the bucket regularly and also removing anything more than 1 month old.
It was agreed that Iain should amend the relevant header files to make the platspec entries explicitly SL5.
Gordon informed the meeting that our older printers (mainly 4100s and 4300s) can be overwhelmed if too many jobs are sent to them within a short period of time. Newer printers are less susceptible to this. It was agreed that we should investigate means of improving feedback to users, particularly by email, in the event of a failed job.
Alastair asked that we should also look at how well the MFDs are handling PDF files and look at improving the feedback should these jobs fail.
Graham stated that the difficulties experienced with the AFS component were due to his investigations into the firewall issue. He had fixed the problem before the meeting.
Stephen pointed out that it is possible to stop the AFS client on a machine without stopping the AFS component.
Topics for discussion
Nothing to discuss.
Alison pointed out that on occasions when the Forum support desk is closed due to the noise from events, the support phone number is forwarded to Ewan or Jennifer's phone as appropriate.
Now that we seem to be close to being able to buy servers again, Alastair asked all units to start thinking about their future server requirements.
The next meeting will be on Wednesday 26th May in room 4.31 chaired by Craig Strachan.
Please contact us with any comments or corrections.
Unless explicitly stated otherwise, all material is copyright The University of Edinburgh