Service Update, March 2018 / Diweddariad Gwasanaeth, Mawrth 2018
The following email was dispatched to active registered users on 7th March 2018:
Dear Colleagues
Following up from previous email updates (https://goo.gl/cy1uv3), we are now in the final stages of preparation to accommodate the Phase 1 system installation in mid-March 2018.
Based on current timetables, Supercomputing Wales (SCW) and ARCCA staff will need to start decommissioning the Sandy Bridge and some of the Westmere compute nodes from the current HPC Wales (HPCW) Cardiff Hub in order to release the necessary rack space in the Cardiff University Redwood datacentre. This work is scheduled to start on MONDAY 12th MARCH 2018. At this point, there will be a reduced service operating on the Cardiff HPCW systems (please note, there should be no impact on the HPCW Swansea or Bangor systems, or Cardiff University’s Raven supercomputer).
Given the decommissioning will involve multiple racks and cables being removed, the HPCW Cardiff Hub should be considered as operating AT RISK during this process. Whilst all endeavours will be taken to minimise disruption, there might be brief outages at short-to-little notice.
In advance of this, researchers are requested to ensure they have copies of all important data currently residing on the Cardiff Hub system – although there is some transitionary maintenance for core services due to the age of the existing infrastructure, should a fault develop on the storage subsystems, we have no guarantee that the data will be fully recoverable. For HPCW systems, there are limited backup capabilities, therefore researchers are actively encouraged to ensure they have copies of all data on other devices.
The new systems are provisionally scheduled for delivery on site by the end of March.
If you do have any queries about the processes and system access/availability or would like support regarding your data migration to other devices, please contact either email address enquiries@
Kind regards
The ARCCA / Supercomputing Wales Teams
Annwyl Gydweithwyr
Yn dilyn yr ebyst blaenorol (https://goo.gl/cy1uv3), rydym bellach yng nghamau terfynol y gwaith o baratoi ar gyfer Cam 1 y broses o osod y system yng nghanol mis Mawrth 2018.
Yn seiliedig ar yr amserlenni presennol, bydd angen i staff Uwchgyfrifiadura Cymru ac ARCCA ddechrau digomisiynu Sandy Bridge a rhai o’r nodau cyfrifiadu Westmere o Ganolfan Caerdydd HPC Cymru er mwyn rhyddhau’r gofod rhestlau sydd ei hangen yng nghanolfan data Redwood Prifysgol Caerdydd. Disgwylir y bydd y gwaith yn dechrau DDYDD LLUN 12 Mawrth 2018. Ar hyn o bryd, bydd gwasanaeth llai yn gweithredu ar systemau HPC Cymru Caerdydd (noder na ddylai hyn gael effaith ar systemau Abertawe neu Fangor HPCW, nac ar uwch gyfrifiadur Raven Prifysgol Caerdydd).
O ystyried y ffaith y bydd digomisiynu’n cynnwys tynnu ceblau a mwy nag un rhestl, dylid ystyried Canolfan HPC Cymru Caerdydd fel gwasanaeth sy’n gweithredu â statws MEWN PERYGL yn ystod y broses hon. Er y bydd pob ymdrech yn cael ei wneud i leihau’r tarfu, gallai fod rhywfaint o darfu dros dro ar fyr rybudd neu heb rybudd.
Cyn hyn, gofynnir i ymchwilwyr wneud yn siŵr bod ganddynt gopïau caled o’r holl ddata pwysig sydd ar system Canolfan Caerdydd ar hyn o bryd – er bod rhywfaint o waith cynnal ar gyfer gwasanaethau craidd yn ystod y broses bontio oherwydd oedran y seilwaith presennol, os bydd gwall yn datblygu ar yr is-systemau storio, ni allwn warantu y bydd modd adfer y data. Ar gyfer systemau HPC Cymru, mae rhywfaint o allu i wneud copi wrth gefn, felly anogir ymchwilwyr i wneud yn siŵr bod ganddynt gopïau o’r holl ddata ar ddyfeisiau eraill.
Mae cyfnod arfaethedig wedi’i drefnu ar gyfer dosbarthu’r systemau newydd i’r safle, sef erbyn diwedd mis Mawrth.
Os oes gennych ymholiadau am y prosesau ac argaeledd/mynediad at y system, neu os hoffech gymorth i fudo eich data i ddyfeisiau eraill, cysylltwch naill ai â ymholiadau@uwchgyfrifiadura.
Cofion cynnes
Tîm ARCCA / Uwchgyfrifiadura Cymru