MPU Meeting Tuesday 18th April 2017


Location sync
The location sync code has been modified to handle George's new data format. Handling location properly was a block to deployment, so that block has now gone.
Location sync test suite
This work has prompted the creation of a test suite, which will attempt to catch all awkward edge cases concerned with location change. For example, it turns out to be troublesome to change an item's location and MAC address on the same day.
It looks feasible to deploy the system this summer. However more testing and documentation will be needed first.
Exam lock scripts and minv
The examlock scripts use minv to get a list of the machines in each room. Stephen tried using ii for this but it seems difficult to pull just the one required field from the output - a custom Perl script would be required. ii can produce CSV output but this data cannot be split on commas (without using a full CSV parser) because they occur in the data. It would be nice to be able to specify exactly what data field(s) ii should produce. Alastair said that it wouldn't be possible to add this feature to the first deployed version of the new inventory system, because deployment is close.


LCFG bugzilla
Chris has skinned up the test server in LCFG livery. Helpful folk in other Schools have confirmed that iFriend access works for them. It's time to deploy the new system.
Final report
It's here. We should all contribute to it.

More DICE disk encryption

Nothing this week.

LCFG client refactoring

Core C libraries
These are now packaged, documented and finished. Stephen has also added some useful example code.
Resource and component libraries
These are packaged. Stephen is working on their documentation.
Perl interface
The package support has been finished. qxpack has been rewritten to use the new API. It works well, so the new API seems solid, appropriate and usable.
Stephen has been unpicking the current arrangement of LCFG packages, moving scripts such as qxpack from one package to another, so that the scripts are grouped according to their relationship to the client data.
Perl XS hack eliminated
Stephen has managed to eliminate a hack which was in use for some years. When creating a new object in Perl XS (which connects Perl and C layers), and the object failed to be created in the C layer for some reason, it seemed necessary to create a dummy object before going on to delete it again. This dummy object is no longer necessary.
Library reorganisation
Stephen is trying to minimise future porting effort by merging multiple libraries into three coherent groups - a core C library, core Perl modules, and client Perl modules. These will be deployed alongside the old ones for a while, until the old ones are no longer needed.

Miscellaneous development

Performance of LCFG slave servers
To try to get our new LCFG slave servers altair and vega to compile profiles as fast as test server muro we've tried disabling hyperthreading on altair. Doing this doesn't seem to make that much difference, but we'll see what effect it has on the time taken to process this week's stable release. We also tried using a ramdisk instead of a physical disk but that made no difference to processing time - perhaps because the test machines have plenty of memory and don't need to swap.
OpenAFS update
An update is imminent. We should deploy it soon. Stephen will schedule this.


IBM array problem
The MPU's IBM SAN box had a problem which Alastair traced to one of its network interfaces. For details see the problem log at the bottom of MPUStorageIBMDS3524.
Root passwords
Chris has been calculating entropy levels and has proposed a topic for discussion at the next operational meeting.
Firewall hole reports
Would we like an MPU-specific firewall holes report? We think we would. What would it look like? Chris will talk to Alison.
KVM jig and rejig
Chris has migrated all VMs from oyster, moved it to AT, repartitioned, reinstalled and tested it, updated its firmware, migrated all VMs from waterloo to oyster, repartitioned, reinstalled and tested waterloo, updated its firmware, and relocated a number of formerly KB-based MPU VMs to waterloo as part of the preparation for the reinstalls of the KB KVM servers amarela and vermelha. We aim to have the KB KVM servers done by the end of May.

Next Week

Next week's meeting, at 15:30 on Tuesday, will concentrate on our RT tickets.

This Week

  • Alastair
    • Inventory project
      • continue working through TartarusWorkFlow
      • Document clientreport (eg how to add modules)
      • Document order sync code
      • Document hpreport processing script
      • Continue work on RESTful API - TartarusRESTAPI
      • Document REST API
      • Write more of the ii commands and document as writing.
      • Test George's new location feed
      • Start work on final report!
      • Convert from mod-auth_kerb to mod-auth_gssapi (See Stephen for details)
      • How represent VMs
      • Continue with REST API testing framework
      • Modify ii query to allow someone just to query one column (eg ii query --location FH-A.30 --field hostname)
    • Deploy encrypted /tmp and swap conversion script
      • Need to warn users that Gnome3 may pop up a window about /tmp being full (when script is run)
    • Schedule MPU meeting to discuss systemd ordering
    • submit polkit bug to redhat - with Stephen (check with 7.3)
    • Think how to regularly report on machines with no /var/log/journal
    • Check sysmans (et al) have 'nograce'.
    • Take a look at RT #78875
    • Look at RT and SL7RT
    • Look at /etc/hosts - dns issue
      • work out what we need to fix current problem
    • Circulate info on RH7.3 systemd changes we may wish to consider
    • Produce a report on machines with wrong time (using clientreport)
      • module written - will hit stable on Wed 29th March
    • Contribute to MPU SL7 final project report
    • Continue looking at LCFG slaves performance
    • MPU Review percentages on PD/OPS/Projects for T1 so far
    • Put sticker on IBMDS3524 to mark controller A FC port 3 as dodgy

  • Chris
    • Inventory project
      • Continue work on clientreport modules for replacing firmwarereport
    • MPU SL7
      • Continue with bugzilla
    • DICE encryption
      • Continue thinking and researching
    • Roll out fixed sleep code
    • Reschedule MPU futures meeting
    • Look at RT and SL7RT
    • Think about whether we can use NX service for staff.login/student.login
    • Produce PXE boot image so that we can update the BIOS of HP 800 G2s
    • Contribute to MPU SL7 final project report
    • Next KVM shuffle
    • Chris to discuss Firewall hole automation MPU specific report with Alison
    • Implement new 'root' password once discussed at Ops

  • Stephen
    • LCFG client refactor stage 1
      • schedule debrief meeting
    • LCFG client refactor stage 2
      • testing and documentation
      • blog article (once documentation complete)
      • complete package support
    • LCFG server symlink to exam branches - produce reporting script and discuss with Graham
    • submit polkit bug to redhat - with Alastair (check under 7.3)
    • Investigate George's multiple network interfaces SL7 issue (eg consoles server)
      • waiting on George breaking metropolitan
    • Look at RT and SL7RT
    • Think about whether we can use NX service for staff.login/student.login
    • Draft a position note on shell components under SL8 and possible ways forward
    • Produce some text for systemd mount bug (to submit to RH)
    • Start SL7 final project report
    • Contribute to MPU SL7 final project report
    • Machine moves (piccadilly, new LCFG slave)
    • Continue looking at LCFG slave performance
Topic revision: r12 - 24 Sep 2019 - 13:50:24 - AlastairScobie
DICE.MPunitMeeting20170418 moved from DICE.MPunitMeeting20170419 on 19 Apr 2017 - 13:29 by AlastairScobie - put it back
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