In the absence of any update, we presume this work is progressing. The temperature went up a few degrees a couple of weeks or so ago, but has since been stable.
Work is well under way to replace the shelves in the Forum self-managed server room with two (initially) more racks. We'll also add one more switch (an ex-closet 2900).
The shelves have been temporarily parked in the main server room, pending a decision as to whether we still need them for chaperoned access to self-managed non-rack machines. (And if we do, whether to move the cluster to the back of the room to make space by the door for the shelves; or alternatively replace the "cluster dexion" with the shelves.)
Chris Adie writes: "A recent Health & Safety incident in the Appleton Tower data-centre has revealed some confusion about the responsibility for IS data-centres. I would therefore like to clarify the arrangements for the management of the data-centres (formerly managed by Alan Reid in LTSTS).
"Day-to-day management of these areas is now the responsibility of Fiona Donaldson (Technical Services Team Leader in CIS). Please cooperate with Fiona and her team in keeping these areas safe, clean, tidy and secure. Please notify Fiona of any potential health and safety hazards you find.
"If an incident should occur, please report the incident to your line manager and to Fiona (in her absence to Chris Adie; in his absence to Tony Weir).
"All accidents and incidents must also be reported to the Director of Health and Safety as soon as possible after the event, and in any case within seven days, by using the University's electronic reporting system at: accident-form
"Accidents / incidents of a serious nature must be notified immediately by telephone (514255) to the Health and Safety Department.
"Specifically in Appleton Tower, there is an issue with the commando connectors (which may be live) dangling too low. IS staff are in touch with E&B, and while we believe the situation has been resolved by E&B, we are currently waiting for confirmation.
"Until it is fully resolved please be aware and extra vigilant."
Incidents involving Informatics staff should also be reported to CEG and to Dave H (with his H&S hat on).
There is one subnet/VLAN in AT for managed machines ("wire M", 129.215.212/24) and one subnet/VLAN for self-managed machines ("wire T", 129.215.2/24). At present the latter has both static addresses and dynamic addresses on it, though only static addresses can have firewall holes opened for them.
Reminder: neither the edge filter holes nor the DHCP servers automatically react to changes of IP address. They only notice lcfg changes, unless you specifically kick them:
om otaka.dhcpd configure
om otaka.iptables configure
The BBC will be out on the back grass again this year. For their comms, they're running two separate fibres through our single-mode infrastructure in the Forum and AT, terminating ultimately in "server room 1" in AT basement. We have one pair patched through from SR1 to B.01, thence to B.02, to one of our closets (in due course), and out of the building to the site. The other pair goes via our AT comms area directly to B.02, and then again via one of our closets to the site.
We're just providing dark single-mode pairs. We have no interest in what's being run over them.
Please contact us with any comments or corrections.
Unless explicitly stated otherwise, all material is copyright The University of Edinburgh