Hi everyone,
Just a quick note that we had a kick-off meeting. I wrote the minutes as far as I understood them. ACTION's are marked using ACTION(name): â
If anything is wrong, then do let me know. If you want to add extra things, do let me know
Regards
Lars
== Attendees ==
Martin Zobel-Helas
Felix Geyer
Kevin Nowara
Michael Sprengel
Ian Jackson
Lars Kurth
= Agenda =
Agree follow-up call for next week to discuss technical details (next Thursday 15:00 - 16:00 UTC)
Should do phone call with OSCAR tomorrow, FELIX will be on the call. Thursday 15:00 - 16:00 UTC (use the same dial-in, if possible)
== Physical access to COLO (Lars)
* Need list of people from Creative (+ID details, passwords, ...) to arrange access to Marlborough COLO
** Access SLA is 24h, but datacenter is normally faster than that
** Sent document template for COLO access to Oscar Herrera
** ACTION (Martin): to chase Oscar Herrera
== Remote Login to machines and VMs ==
* ACTION: Martin to send ...
** Send a mail with all the PGP public keys
** Plus, who keys belong to, etc.
** Ian: Could do Ansible for access, also have no objection re puppet, but have no experience with it
** ACTION (Ian): share COLO details in Marlborough with Martin and set up access for Rackspace VM panels
* Ian: Do not have a shared password system (about 10 different passwords)
** ACTION (Martin, Felix): decide what the best way forward is (e.g. can were use Credativ's system)
Ian Jackson Overview
* Two completely separate areas
** Web services (git, wiki, mailing lists, blog) ... runs on Rackspace VMs
** Test Lab is hosted on Rack in Marlaborough data centre
== Communication (Agreed) ==
* Regular calls at the beginning : twice a month - for status update and open questions - one call for everyone. May not need calls that frequently. Can keep calls in calendar and cancel as needed.
* For official things use RC tickets for SLAs and time tracking
* Use IRC (aka #xendevel on freenode) for quick unofficial questions, etc.
** Could also define #xeninfra
* For minutes: use test wg mailing list (can sign up Credativ staff to it)
** ACTION(Martin): get back to Lars with Contacts
== Workflow ==
For tickets and communication, Credativ uses RequestChecker (RC) for all customers
* ACTION: Share ticketing system info with Lars and Ian
For pro-active work: use the same approach and use
* ACTION (Lars, Ian): provide list of people (emails) who can raise tickets against credativ
Martin: does some sort of Admin VM exist for the COLO?
Ian: There is a management VM in the COLO for DNS, etc (not sure I understood this correctly)
Ian: We do have the capability to create new VMs in Rackspace
Ian: Have a few "empty" VMs which could be re-purposed or retired
= Starting Tasks =
Agreement: set channels first, then stuff via tickets. But come up with a rough list for planning ...
Agreement: agree on a priority list on next call
== Issues we know ==
* Lacking back-ups and monitoring for Debian machines in COLO and on Backspace VMs
* Not clear what VMs are up-to-date with regards to latest security patches
* Some HW issues with COLO
** ACTION(Ian) to re-send list once ticketing system is in place and provide pointers to all necessary information
** 3-4 ARM boards which lost SATA disks
** 4 x86 test boxes which have something wrong with them
* Lower priority items:
** Sysadmin management for Rackspace VMs
