Re: [Wg-test-framework] Notes from Xen Project/AllNet call Mar 4th

[Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>]

Attendees
---------

Paul George
Don Koch
Ian Jackson


Overall
-------

Don: We have 8 x86 machines outstanding.  [ By implication the other
16 are all fully set up. ]

Ian: Pleased that we are making good progress now.

Ian: Is now unblocked and working on software setup.

Ian: Most worried about ARM crate, as that has most yet to do and
therefore greatest opportunity for unexpected problems.


ACTION: Don to send his status/checklist to the wg list.

Paul/Don: There are updates to the colo layout (wiring assignment)
spreadsheet.

ACTION: Don to send latest sheet from git to Paul for updates;
ACTION: Paul to update it and send back to Don
ACTION: Don to commit and push to git


x86 machines status report
--------------------------

Don: We have 8 machines which are not entirely ready:

 4 need more serial adapters

    Paul: Chris has probably made some of these, but is not sure.

ACTION: Paul: have Chris make more if necessary, and fit.

 2 "rimava" outstanding AMD machine MSI serial BIOS problem

    Last meeting: Paul was going to look for a replacment m/b

    Paul: His supplier is looking for a suitable board, either a
    server board with an A3 socket or a workstation board with serial
    BIOS access.

ACTION: Paul to continue to pursue this

 2 "huxelrebe" do not appear to be able to do BIOS over serial

    Don: Connected VGA console and went through the BIOS screens and
    did not see this option

    Paul: This is strange because Paul bench-tested each machine and
    noticed that rimava* didn't seem to support this so would have
    expected to notice this for another machine too.

ACTION: Paul to (liase with Don and) investigate.

Don: Additionally huxelrebe1 does not appear to have an OS install

ACTION: Paul to liase with Don and install OS on huxelrebe1


ARM crate status
----------------

Paul: The crate is in the rack, but not yet connected.  The 4 needed
    serial adapters have been made by Chris and therefore the machine
    can be connected this afternoon.

Ian: Gave overview of ARM crate.
There are the following connections to make:
     * 4 DB9 serial connectors (arndales)
     * 9 RJ45 ethernet connectors: one for each board (to
       test network) plus one for built-in PDU (to infra network)
     * 1 USB connector for built-in serial for cubietrucks
     * IEC mains power
[ Also: 8 DC power leads looped back to boards from inside crate. ]

ACTION: Paul to connect the ARM crate this afternoon

Paul: will need IP address assigned for PDU

ACTION: Paul to provide Don with git URL for installation instructions

ACTION: Don to set up PDU MAC address with an IP address

ACTION: Don to allocate addresses for 8 ARM dev boards (MAC addresses
in the ARM installation instructions git repo), ask Ian if not found

ACTION: Paul to test the ARM crate in situ according to existing test
plan

Paul asked about tty number assignment:

Ian: tty number assignment is not at all stable; need to redo the grep
each time.  (For production we will configure server specially but
this is impractical for testing.)

Paul: grep did not appear to work, but lines were found in logfile.

ACTION: Don to assist Paul with this grep


ARM crate USB serial passthrough
--------------------------------

Ian: We want to pass USB serial connections for cubietrucks through to
serial0 VM.  This would be easiest by doing PCI passthrough of the
relevant USB controller.  Ian is not convinced about stability of USB
passthrough.

ACTION: Don to investigate which physical USB ports on newcastle
correspond to which USB controllers visible in lspci and lsusb, by
using physical USB keyboard (already present in rack) and ARM crate
cubietruck serial.

ACTION: Don to discuss results and construct plan with Ian.

Ian: This may involve physically labelling some of the USB ports on
newcastle to avoid confusion in the future.


Liason about sysadmin and hardware fiddling
-------------------------------------------

As between Paul (hardware) and Don/Ian (software), all physical x86
machines now "owned" by Don/Ian, except where agreed physical work is
taking place.

As between Don and Ian: Will negotiate on IRC about disruptive
activities (eg rebooting VMs, etc.).


Thanks,
Ian.

_______________________________________________
Wg-test-framework mailing list
Wg-test-framework@xxxxxxxxxxxxxxxxxxxx
http://lists.xenproject.org/cgi-bin/mailman/listinfo/wg-test-framework