This maintenance will not start until 17 September. Our status page platform has sent out a couple of incorrect emails about the timing of this maintenance; apologies for the confusion.
As previously announced, UIS have notified us that essential electrical works will be carried out at the West Cambridge Data Centre (WCDC) on 18 September 09:00-17:00. The entire data centre will be switched off all day.
This will affect many departmental services (several of our servers are hosted in WCDC) as well as other services in the broader University. Many departmental services will be shut down on the evening of 17 September and will remain offline until at least the evening of 18 September.
More details on the precise impact:
The following departmental systems will be unavailable:
Departmental databases: dbwebserver, svr-win-db
All research and teaching VMs, except dev-gpu-* and dev-cpu-*
SSH servers: ely, svr-ssh-0 (use slogin.cl.cam.ac.uk instead which will be changed to point at svr-ssh-1, hosted elsewhere)
Webadmin
Verex
Archive servers (archive / berilia; archive-smb / jerakeen)
Licence servers: lmserv-*
Cron servers (cron-serv*)
Weather station
Undergraduate SSH servers: cl-student-ssh, cl-teaching-ecad
WSUS
Misc utility websites on svr-www-02, svr-www-03, www-dyn*
Legacy Windows remote desktop service (clrds / desktop)
Legacy Subversion server (svn1)
Legacy printing (CUPS, mDNS/Bonjour)
Legacy VPL servers
Legacy wiki server
Morello: entire cluster
EEG: beara, gola
SRCF: egress, echo, enid
TFC: tfc-app{1,2,4,10,11}
UIS have told us that the following University-wide services will be down all day on 18 September:
The following departmental systems will be operating without resilience and are at risk of disruption:
VPN2
Active Directory (adsrv07)
DNS (authoritative and recursive)
LDAP
DHCP
MSA, MTA (SMTP outbound)
TGT servers
Legacy email (on filer / using .forward files)
cl-onserver / laira, march
Filer disaster-recovery snapshots
The following departmental systems will have their outage mitigated, i.e. they will be rehomed in the William Gates Building prior to the outage: