MPU Meeting Tuesday 30th March 2010

AFS Component

Nothing happened, waiting for sign-off.

LCFG Server Refactoring

Nothing much happened.

Server Hardware

Stalled.

Installroot

Nothing much happened.

F12

There's been lots of progress. Of the outstanding items on the devproj list:

• lcfg-syslog is done, but check whether the platforms list has been updated before striking it off.
• lcfg-openafs is OK on 64 bit. Chris will check it on 32 bit. For this he'll need Stephen's latest version of the perl-AFS module, which Stephen will submit.
• Talking of perl-AFS, there's a new release of it in preparation. The author has Stephen's patches, and even better, he's aiming for 64 bit support.
• Iain is working on lcfg-pam.
• Chris should delete lcfg-rpmcache and lcfg-xinetd from the devproj list.
• Stephen will tackle lcfg-cyrussasl and lcfg-gdm.
• Alastair will do lcfg-xfree but not yet; it's not needed for straightforward desktop machines.
• On the subject of lcfg-xfree, we currently use it to set alternative keyboard layouts. A better way to do this would be to get lcfg-hardware to set the keyboard layout using /etc/sysconfig/keyboard as X now respects that. Chris has reported this as bug 263.
• Chris will add lcfg-sleep to the "Other components" list.
• Chris will build and submit (to devel) everything else in subversion as yet untackled.
• Alastair has been working on lcfg-fstab and lcfg-hackparts.
• Chris needs to make an fstab template for tarragona based on the current /etc/fstab; see bowmore for clues.
• lcfg-grub works, though splashimage doesn't (bug 256).
• lcfg-hardware works but the platform list needs updating.
• lcfg-install is on Alastair's testing to do list.
• lcfg-kernel works but the platform list needs updating.
• lcfg-network is still to be tackled.
• The installroot has been done. The installbase hasn't been done yet.
• PXE is still to be done.
• The "port MPU managed resources to the DICE level" step consists of:
  • propagate the new inf-level settings to other levels as appropriate
  • Check all LINUX_SL5 conditionals and consider whether or not to extend them to F12 too.
• Chris has made a start on the F12 port page but there's a lot to be summarised! It'll also be used for (e.g.) the hardware test results. We can do the hardware tests once the install tech is fully working.
• Stephen has been using Mock to build everything for f12_64. The dependencies seem very nicely sorted now and the build is proceeding smoothly.

Boot

Alastair will finalise his boot work this week.

Miscellaneous Development

Network component improvements

• Bug 66: Stephen has produced a patch; Alastair has looked at it and will comment on it; Stephen and Alastair have agreed a solution which Stephen will implement.
• Bug 62 (doesn't warn if address lookup for interface fails) and
• bug 103 (allows whitespace where it shouldn't) also need fixing.

Remote hardware management for desktops

Stephen has added this new project proposal.

DICE F12 Desktop

Alastair has added this new project proposal.

SEE project cooperation

Alastair spoke to Colin Higgs. They'll be happy to help with the SL6 port but they're not sure that they'll be using intervening desktop solutions such as the F12 one.

LCFG template bug

Stephen found a bug (bug 259) in the LCFG template handling: when creating a new file it doesn't allow restrictive file permissions. Stephen has fixed the template handling so that it now honours the calling environment's umask setting.

Operational

disk space for telford

Gordon's looking into it for us. Craig would like us to use smaller AFS volumes if possible, and Stephen has spotted that our F12 Everything mirror changes a lot less often than our F12 updates mirror - so Stephen suggests that when we get the disk space, we put Everything and updates into separate volumes. The smaller number of files should then make volume releases much quicker.

Wire 212 servers

mousa has been moved off the 212 wire in AT. split is still to be moved.

Sleep user education

Chris has produced some illustrated posters and stickers which should educate student lab users about how to wake a sleeping machine successfully and non-destructively.

cron log file

Alastair has changed cron logging so that it logs to /var/lcfg/log/cron rather than /var/lcfg/log/syslog. Permissions on both files have also been tightened up.

mrtg

Alastair has removed it. It did nothing but clutter our log files.

This Week

Alastair will:

• MPU disaster recovery
• Look at Panos's lcfg-network change
• MPU logging notes for George
• Talk to Craig about bpbeast
• Installroot

Chris will:

• MPU disaster recovery
• F12

Stephen will:

• MPU disaster recovery
• F12/x86_64
• Change the IP address for split
• om changes
• lcfg-server update

-- ChrisCooke - 30 Mar 2010