MPU Meeting Monday 25th February 2013

Server Upgrades

No change. Stephen needs to add his notes to the final report so it can be submitted for sign-off.

Security Enhancements

Nothing happened. Stephen plans to prioritise finishing the documentation.


Stalled, might be able to restart this project in the next week or so once the dust has settled on the new docproj service.

Login Logs Viewer

Nothing happened. Need to chase George about the status of the Data Protection statement.

Sleep Enhancements

The component now detects X idleness. It has a new resource which is used to specify a time delay between X becoming idle and the machine being considered ready to sleep. This avoids the problem of machines going to sleep immediately after the screensaver has been activated.

Chris now plans to widen testing beyond COs to a few well-selected real users.

Miscellaneous Development

Various chunks of code have been reworked so that a problem can be caught and an Error message issued rather than just calling Fail. This means that the component can keep going in odd situations which previously prevented it from starting (and thus being usable to fix the issue later). Closes: #555. This problem is almost certainly very widespread, at some point we might want to check through some of the other core components.

Added support for setting the From and Reply-To headers when sending an autoreboot message by email. Closes: #635

Added support on Linux for configuring the portreserve daemon and running the portrelease tool. The schema will be rolled out first and then extensive testing done before the new versions of ngeneric and LCFG::Component are deployed to the stable release.

A new component has been developed for managing access.conf style files. At this point we do not plan to use it to replace the basic functionality supported by the auth component for managing /etc/security/access.conf via the auth.users resource. We might want to remove that support when we do the next platform upgrade (e.g. SL7). For now we will just use the accessconf component for more complex authorization schemes.

The trigger support is now working properly and being used to rebuild the initramfs when the qlogic drivers are installed or upgraded.

LCFG header refactoring
We need to discuss within Informatics. We will raise the issue at the next Operational Meeting and ask people to consider the topic for the following meeting.

EPEL Updates List
Michael Gordon suggests that it would be good to have a list which automatically updated the versions of packages in the EPEL repository as happens with packages from SL. It should be easy to generate this list using the current package list scripts but we will need to make it clear that the list is not really "supported".

CPP Expansion
Michael Gordon suggests that cpp expansion, except for mutations, done on the server would be a useful optional debugging tool. Kenny added that this might be requested by a pragma at the top of the source file. We think the best way to do this would be to add a debug script (a bit like dumpdeps) which can be run on the server using the LCFG server resources to set the correct include paths, etc. We don't have time to do this so we will ask Michael if he could contribute a script.


openafs 1.6.2
This is close to being released. It contains a couple of security fixes and we should aim to get it onto the fileservers fairly quickly once it is released.

RHEL 6.4
This has been released. We expect to see SL6.4 in a month or so, in the meantime there will be lots of updates backported to SL6.3. We will probably be deploying this at the end of May after the end of teaching.

We would really like to remove support for SL6.2, Chris will chase this up at the Operational Meeting.

Daily KVM report
This now lists volumes which appear to be unattached. It now warns when the current pool is wrong so the machine can be scheduled for a power-cycle. It also shows the cputime.

DPMS settings
Alastair has discovered the gconf settings that need to be tweaked to control when the screen is blanked after the machine goes to sleep. This could be reduced considerably.

RT bugs
We need to look through our RT queue and respond or close tickets where necessary.

This Week

  • Alastair
    • Check whether console wire is accessible on DICE desktops In principle, but George not happy. Perhaps use web browser on marriner. George thinking about
    • Educate individuals about inappropriate KVM guest sizes
    • Look at gconf component for reducing default time for monitor display turnoff
    • Create an MPU KVM server header - finish off and deploy
    • Sort out RT bugs
    • Review action list

  • Chris
    • Take SL6 project for signoff
    • Sort out RT bugs
    • T1 figures
    • Review action list
    • More testing of sleep project
    • Start off Lab exam virtual image (flesh out project plan and take to development)

  • Stephen
    • writeup buzzsaw documentation
    • flesh out SL6 project report
    • Sort out RT bugs
    • T1 figures
    • Review action list

  • Carol
    • carry out an audit to make sure that all guests are still required
      • a) to free up resources for new guests,
      • b) shouldn't leave unused, half-managed machines lying around

-- AlastairScobie - 25 Feb 2013

Topic revision: r5 - 05 Mar 2013 - 09:27:13 - StephenQuinney
This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback
This Wiki uses Cookies