Services Unit July 2019 Power Work

Things we want to do while the power is off, or just after it comes back, to the Forum Server Room.

  • move gresley (to rack 4) and peppercorn (down a U) (to leave space for possible riddles JBOD) - ForumRackPopulation
  • install JBODS below gresley and peppercorn
  • attached 4TB JBOD to ifevo2, probably shuffle boxes to make cabling neater - ATABoxesInfo
  • kraken is reporting a fan 3 has failed. Swap one from nessie or yeti.
  • upgrade afsdb from AFS 1.6 to 1.8 - It will be a version of the last UpgradeAFSDBPlan
  • SL7.6 all off machines - probably switch profiles on the Friday so they only have the onboot RPMs to do on restart.

To do before the power off, probably Friday afternoon.

  • Switch printing to borges - DONE
  • Switch www-dr to be the live www.inf - ServiceUnitWwwOffsiteDrPlan (notify any recent CVS publishers)
  • Remove the #define DICE_STICK_WITH_SL75 from the list of machines below. LcfgStop is probably useful - DONE

List of Forum machines still on SL7.5

arachne         catfish         dunkvm          holler          skelp
babble          chatty          edinnlpvm       huldra          stoatervm
bahookie        cigarvm         fanklevm        kegansl7        tarnvm
banter          ckankvm         flakey          kraken          tench
blabber         collins         gabble          lardup          windydev
blenny          cory            gorgon          mcgraw          windy
blether         craic           gresley         peppercorn      wishart
blogvm          crivvens        gype            prattle
breezy          croxley         hanlon          ranter
buckrogers      danio           harpersl7       rhacotis
calvinvm        drip            haver           riddles

After the event notes

  • ifevo2 needed fixing following shuffle and attachement of extra jbod
  • crivvens (bpweb) /home/UUN stopped working
  • ranter (jabber) room service doesn't start automatically, manually OK
  • gresley a RAID disk member failed
  • Need to tidy/fix printer dhcp/gateway - should use infrastructure machines for DHCP and routing, so no network dependency on cups server
  • general lack of some IPv6
  • huldra tibs DRIVE IDENT error
  • quotas needed restarted on gype and bahookie (they powered up before quotas host)
  • some servers needed extra reboot as they were up before package server was responding (ie SL7.6 didn't happen)
  • buckrogers RPM clash (SL7.6 related)

-- NeilBrown - 18 Jul 2019

Topic revision: r6 - 23 Jul 2019 - 10:59:20 - NeilBrown
 
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