MPU Meeting Tuesday 1st March 2016

Inventory

Nothing happened.

LCFG Client Refactoring

Some work has been done on the package handling, this is now pretty much complete. The LCFG::PkgTools Perl modules have been updated for the latest API, everything is working quite nicely now, using the new interface is much simpler. The current focus of the work is on making everything much more robust, in particular with ensuring that there are no memory leaks when errors occur and that sensible error messages are returned.

SL7 Server Base

network names
The "modern" network interface names have now been set correctly for the desktop models we have available. The Dell T5610 still needs doing, Alastair suggested booting with the SL7.2 install DVD to find out what the interface name will be. We think that the change in interface name will only affect VirtualBox if a bridged adapter is being used and that NAT should be fine.

LVM and multipath
Alastair has finished the work on multipath and LVM, he will write up the details as a blog article. It appears that dmraid is installed and enabled by default. It is only required for software RAID, which we don't use, and it causes a lot of noise so we plan to remove it and provide an options header in case anyone wants it re-enabled. There are many parameter changes for multipath. Alastair thinks we need to look at the configuration for the Dothill arrays on SL6. It appears that we can now avoid configuring multipath and LVM in the initramfs, by default it is not pulled in. This is good as it means we can avoid having to rebuild the initramfs for each change. There are new resources in the multipath and LVM components to control whether or not a rebuild is required. This has all been tested but maybe we have just been fortunate with the timing? It looks like the problems with pvscan were caused by lvmetad. That appears to be unnecessary so we now turn it off. Just waiting on Matthew Richardson checking the patches and giving us the nod, we will chase at the next LCFG Deployers Meeting.

localhome
The localhome component has been updated to support autofs as well as amd. The various autofs configuration sections in the dice headers now honour the AMD macros. Ideally we would rename the macro to something more generic.

MPU SL7

NX service
There is an SL7 version of the NX service running on sauce. It would be helpful if everyone could test this from home. We need to decide what to do about the broken support for Gnome.

SSH service
There is an SL7 version of the SSH service running on rabbit. This is basically ready to go but we still need to check the auditd reports. We will keep the test SL6 SSH host - shrew - until we have upgraded both servers. We should do the staff host first as users who are allowed to access that host could always use the other if any problems are found.

apacheconf

The work on the apacheconf component for SL7 and apache 2.4 has been completed. Stephen needs to give a talk on this at a Development Meeting.

Miscellaneous Development

hackparts & fstab
Support has been added to the hackparts component for setting disk flags. This is needed for the new HP G2 which requires the pmbr_boot flag be set.

disk encryption
The script for enabling encryption of /tmp and swap now seems to work well. We will deploy to office desktops soon but will leave the student labs until the end of teaching.

rkhunter
The rkhunter component and configuration have been updated for both SL6 and SL7. The daily reports now work correctly again.

SL7.2
New users and groups were added for the geoclue2 and libvirt packages. The default kernel has been updated 3.10.0-327.10.1.el7. A couple of package versions were fixed in the installroot and installbase package lists as they were still on the pre-release versions.

LCFG server
The server has been updated to version 3.4.2 which adds the new mIFNULL mutator and also adds support for the newline translation in resource values.

SL6 single user
The default behaviour for SL6 single user mode was inadvertantly changed from sulogin to sushell when we fixed up the scripts a while back. This meant that there was no root password prompt. The previous situation has now been restored.

Operational

Reboots
The SSH and NX servers were rebooted for the glibc security update.

waterloo
A failing disk was reported on waterloo, the disk will be replaced. We should also schedule downtime for an upgrade of the various firmware.

firmware updates
Chris will download the latest firmware for the Dell R710. Stephen will update the firmware on circle, Alastair will update metropolitan.

polkit
We have found a problem with the version of polkit in SL7. This causes general protection faults. On most machines this is not fatal but for some reason the Dell T5610 theia was regularly hanging. We think it might be related to the AMD Catalyst Pro driver it uses. The fix is to upgrade from 0.112 to 0.113 which is in Fedora. Sadly this isn't simple since the i686 version is not buildable, to workaround this the i686 version of the GConf2 package must be removed. This is in the redhat bugzilla as bug#910262, we should chase this up with Redhat. We wonder if this might be the cause of other machines hanging? Stephen will investigate.

computing.help
Are there are MPU computing.help pages will stil need reviewing?

This Week

  • Alastair
    • Inventory project
      • continue working through TartarusWorkFlow
      • consider what next can be integrated into existing system, if anything
      • Check for systemic errors from clientreport
        • Look now that servers don't check monitors
      • Document clientreport
      • Document order sync code
      • Continue work on hpreport processing script
      • Consider milestones for inventory project
    • Remove default pool if ops meeting agrees
    • Think of a use for 'atom'
    • Deploy encrypted /tmp and swap conversion script
      • Deploy on office desktops now
      • Leave labs until end of teaching.
      • Need to warn users that Gnome3 may pop up a window about /tmp being full (when script is run)
    • SL7 base server
      • Rename the AMD macro that people use to enable/disable automounter on machines with localhome
      • check metropolitan USB and CD
      • Continue work with FC and LVM
        • await matthew apply patch to lcfg-lvm and deploy
        • finish draft blog article
        • check nagios notices if FC cable removed
      • network and bonding
    • Schedule MPU meeting to discuss systemd ordering
    • Continue building computing.help honeypot
    • package up ILW stuff and document process
    • Read through SL7.2 release notes
    • Dig out SL7.2 live CD (for Stephen)
    • Install 'opennx' client on SL7 (currently manually added to gala)
    • Apply firmware patches to metropolitan
    • submit polkit bug to redhat - with Stephen
    • computing.help - email on feedback page -> new RT ticket email address
    • Check device naming for virtualbox and kvm guests under consistent naming schemekvm = eth0, virtualbox=en0ps3 (but that depends on exact virtualbox config which is under control of user, so can't generalise

  • Chris
    • Inventory project
      • continue working through TartarusWorkFlow
      • Look at clientreport modules for replacing firmwarereport
    • pkgsearch for SL7
      • reimplement as a yum web front end (yum search for keyword produce an html file of links to cgi to do yum info)
      • Need support multiple platforms
    • Liaise with George over iDRAC documentation (look through ops reports to remind)
    • SL7 base servers
      • test out rsync / rmirror (both client and server ends) - liaise with Neil
    • Continue investigating SL6 sleep problem
    • Schedule MPU stargazing meeting
    • Try looking at ipmi-sensors on "theia" (suspected fan issue)
    • Download R710 firmware patches

  • Stephen
    • LCFG client refactor stage 1
      • schedule debrief meeting
    • LCFG client refactor stage 2
      • document API
      • blog article (once documentation complete)
    • apacheconf
      • produce some recipes
      • talk at devel meeting
    • Think about PD - Interested in ZeroMQ
    • Investigate kernel component pipe moan by using shell commands instead of RPM module => waiting on 7.2 => activities list
    • LCFG server symlink to exam branches - produce reporting script and discuss with Graham
    • Circulate dmesg proposal
    • Apply firmware patches - circle
    • submit polkit bug to redhat - with Alastair

-- AlastairScobie - 01 Mar 2016

Topic revision: r7 - 23 Sep 2019 - 13:33:38 - AlastairScobie
 
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