canonical-ci-engineering team mailing list archive

[Evan Dandrea <evan.dandrea@xxxxxxxxxxxxx>]

---------- Forwarded message ----------
From: Larry Works <larry.works@xxxxxxxxxxxxx>
Date: 16 November 2013 05:32
Subject: Lab update for November 15th
To: Evan Dandrea <evan.dandrea@xxxxxxxxxxxxx>


Evan,

Here where we currently are:

Landing Task Force stuff: I am assuming this is mostly s-jenkins. If
that is indeed the case then there is no major change as yet. The RAID
on naartjie seems to have a failed drive. Fortunately it was a RAID5
configuration so it's mounted in a degraded state at the moment. Rick
install a 10 gigabit NIC in it and I am currently backing it up to the
storage server. Once that is done I will run the VMs from the storage
server to see what kind of performance we get; if it isn't decent then
we will install a replacement drive and go ahead and replace the
controller (since wh just happen to have both on hand), build a new RAID
and transfer the data back. Either way we will have that piece up by
Monday. All of the Calxeda nodes are up but I believe you already
received that information. That will leave just a few desktop systems to
be connected and Rick should be going by MicroCenter to pick up the
pieces we need on his way back from Lex tomorrow.

Kernel team hardware: All of the previous hardware they had should be in
the same operational state it was prior to the move. The new hardware we
racked for them is racked and connected to kvm/network/cdu but we
haven't done anything else with it (configured them for net boooting,
configured them to PXE boot from the correct server (ours or their MaaS
instance). That is a bit too far down the priority ladder at the moment.
The hardware we have for kernel SRU is up and running. The only thing we
will have to work through is getting the jenkins jobs to work with the
hardware being on an APC PDU instead of the normal ServerTech PDUs. The
APC PDUs are network accessible now so that should be doable.

Cloud: I worked with James Page and the HP stack is network accessible
and James has them mostly back up. I don't have any more details than
that. The network accessibility issue wound up being a simple thing; we
hadn't configured the Cisco switches to connect to the HP managed switch
the HP stack is connected to. The iLO issue only involved a few of their
system that, for some reason or another, didn't pick up the default
route from the DHCP server.
James didn't respond to my query about the Openstack test suite but that
was probably because it was getting close to midnight his time and he
probably had better things to do on a Friday night :-)

 The only external MaaS stack interface is up and accessible. Diogo
wasn't online when I checked so I don't know if he has checked his
systems out yet or not. I will get with him on Monday.

OTTO: No changes there yet. As stated, Rick should be picking up the
parts we need to get the rest of their systems back online hopefully by
tomorrow.

In other news I did get most of the m-jenkins nodes connected to their
jenkins master. The remaining three are also waiting parts. The previous
two cobbler instances have been combined into one. And we can definitely
PXE boot now.

I will go through the list of tasks tomorrow morning to see what else I
can get taken care of over the weekend but for now it's Saturday (ableit
very early Saturday) and and I thinkg I'm going to enjoy an adult
beverage and then get a few hours sleep.

~w