MPU Meeting 10th July 2006

Ken's made a prelink component already based on the file component, so Stephen just has to add some headers and add some local DICE exclusions to the list (ACTION: Stephen). Stephen added that prelink seems of much less use than it was because of the address randomisation stuff that's been added to the kernel.
LCFG distribution
It was written up and taken to the development meeting. Now we need to get a detailed proposal together. Stephen will think about it and meet with Paul over the next few weeks (ACTION: Stephen). He'll also put together a page about this and buildtools so he can ask for feedback at the SciCOs meeting tomorrow (ACTION: Stephen).
Chris will check the state of play and how compatible they are (ACTION: Chris). (Minor surgery is needed on dice/os/fc5.h after which we can make a diydice/os/fc5.h which is all we'll really need.)
FC5 on HP d530
Chris will have a go installing FC5 DICE on a d530 and let Stephen know how it goes (ACTION: Chris). (The first attempt fell over with a disk partitioning hiccup but the second attempt succeeded.)
FC5 LCFG components
Stephen has brought the LCFG components up to date on FC5.
Toby's patches
Toby has some outstanding LCFG Mac patches which haven't yet been applied. Paul will check with him (ACTION: Paul).
FC5 Progress
Other units are now putting a lot of software on FC5 DICE. Graham Dutton is installing the Java software.
New status page without Javascript?
Following George's comment, is the LCFG status page usable using the new code when Javascript is disabled? Stephen will try it out (ACTION: Stephen).
Status page hierarchy
Prompted by Lindsey's RT ticket 24830, and the imminent arrival of hierarchical divisions in the LCFG status page, Chris will move us towards having a standard structure for the status page hierarchy (ACTION: Chris).
64 bit progress
Stephen has a client working with a minimal 64bit LCFG installation, which is a big milestone. The next step is to get package support so that LCFG components can be installed automatically rather than manually. All noarch components have been installed. He discovered that some components build as i386 when they really ought to be noarch; he's fixing that. This flags up a more general point: with the arrival of 64bit we'll have to restructure the directories in which we store package files. This ties in with Alastair's revamp of RPM submission.
Achilles disk space
This morning achilles filled up with slapd logs: there were half a dozen of them at up to 2GB each. We should find out what LDAP role achilles is playing and try to get it separated from achilles' basic roles of LCFG master and LCFG subversion server. Chris will talk to Toby about LDAP on achilles (ACTION: Chris). Stephen will look into tweaking logrotate resources to make it gzip older log files (ACTION: Stephen). Stephen will look into tweaking logserver to make it cope with gzipped logs (ACTION: Stephen).
New logserver project proposal
Stephen has been having some interesting thoughts about the status pages and the logserver component which he'll put into a place-holder project proposal on integrating logserver with Cosign (ACTION: Stephen). See:
Craig's Bugzilla proposal
We can't make up our minds on this without Alastair so we'll wait until he comes back.

-- ChrisCooke - 10 Jul 2006

Topic revision: r8 - 12 Sep 2006 - 11:00:37 - ChrisCooke
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