Infrastructure Unit machines

MachineSorted ascending Role Comments
All IT closets   power-down/up
All server-room edge switches   will be turned off as part of general power-down
ancerl AT netServ to be not in service yet - will eventually be the AT equivalent of hickox: VPN endpoint, etc.
atda AT decant switches keep one of these running for jarvi
atdb AT decant switches keep one of these running for jarvi
barrett KDC master remove _kerberos._udp SRV record to speed up authentication; om kerberos push to slaves; take dump of kerberos db
beziers AT consoles (decanted) Is the console server for Forum network machines hogwood, harnoncourt and marriner, (as well as AT machines). Ideally should be last down, and first up.
core0 Forum core switch keep running throughout for AT external link and hogwood
core1 Forum core switch switch off
core2 Forum core switch try to keep running throughout - has link to one of the AT basement switches
core3 Forum core switch turn off? or try to keep running throughout?
darwin AT extNS (decanted)  
fenrir KDC slave/AFSDB down late, up early; remove _kerberos._udp record
franklin LDAP master take dump of ldap db
harnoncourt Forum extRt  
hickox Forum netServ relocate power to netServ UPS
hogwood Forum netInf keep running throughout for Forum monitoring
Hot-spare switches   can be turned off in advance
jarvi AT netInf (decanted) keep running throughout for AT network services
kingsmen toby test machine  
kubelik AT extRt (decanted)  
linnaeus Forum extNS  
marriner Forum consoles Last down, first up - but see beziers!
mckinley LDAP slave move infdir alias; remove _ldap._tcp SRV record
osprey cosign/ifriend KDC master remove from weblogin alias the night before; om kerberos push to slave; take dump of ifriend kerberos db
panther KDC slave/LDAP slave move various ldap aliases; remove _kerberos._udp and _ldap._tcp SRV records
sr11 Not in service yet switch off
srif Forum SRIF PoP switch switch off

Comms rack sequence

Machine Down Up Console Notes
hogwood - - IPMI Keeping up
marriner 7 2 IPMI Console server for most of server room
harnoncourt 2 7 IPMI  
linnaeus 4 5 srslc02  
hickox 6 3 srslc02  
beziers 8 1 none Console server for marriner and harnoncourt
darwin 5 4 none  
jarvi - - IPMI Keeping up
kubelik 3 6 IPMI  
ancerl 1 8 none  

Notes

  1. Power to the main server room will be removed by throwing the breakers in the distribution board. The three for the core switches have been labelled; it'll be interesting to see whether they're correct or not. Restore power after marriner is back up.

  2. Power to the self-managed server room will be removed by throwing the main switch in the distribution board. Restore power after everything else.

  3. Power cables in the comms racks to be labelled to make the necessary disconnections/reconnections easier.
    Done - idurkacz

  4. Trailing TP cables to be set up from core0 for CO's laptops.
    In hand: the eight core0 ports C13, C14, ..., C20 have been set to DHCP; cables need to be attached - idurkacz

  5. The unused Sanbox 5600-2 is to be liberated from Rack 4.

  6. Techs' help to be arranged for powering off the closets.
    In hand

  7. Turn off Nagios alerts for the entire period by:
    1. Just before the power-off, hand-editting the file curlew:/usr/bin/notify-by-jnotify and replacing the line
        my $message = join("\n",<STDIN>);
      
      by
        my $message = join("\n",<STDIN>);
        exit;
      
    2. On curlew, om nagios_server stop; om nagios_server start

    Revert all this after the power-off is over.

-- TobyBlake - 15 Dec 2009 -- GeorgeRoss - 15 Dec 2009

Topic revision: r10 - 15 Jan 2010 - 14:14:59 - IanDurkacz
 
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