On 4 June 2013 21:45, Matthew Gretton-Dann matthew.gretton-dann@linaro.orgwrote:

Dave,

I notice that nodes 11-15 inclusive seem to have disappeared from the calxeda02 box:

Dave installed a new node and some of them need re-configuring. But I'm still running the LLVM release and would love if I could have a few more hours before someone touches the server. We know how sensitive it is and I already ran it three times today.

I'll let you guys know when I'm done with the release, so you can update the configuration.

Thanks, --renato

Something very odd here, post the broken node replacement. Matt and I are investigating and need exclusive access today, because we need to power cycle it, probably more than once.

Thanks

Dave

On 4 Jun 2013, at 21:45, Matthew Gretton-Dann matthew.gretton-dann@linaro.org wrote:

Node 16: 192.168.1.111 Node 17: 192.168.2.126

https://jira.calxeda.com/browse/LINARO-1

Back to Calxeda and ask their input. For the record, in the same ticket Chris Oakley from Calxeda who's following this case had requested that we arrange a chat with him to coordinate the swap out and walk through FW upgrade after, did that chat take place?

Best regards,

-- Ilias Biris - ilias.biris@linaro.org Technical Program Manager, Linaro M: +358504839608, IRC: ibiris, Skype: ilias_biris Linaro.org | Open source software for ARM SoCs

On 5 June 2013 13:10, Dave Pigott dave.pigott@linaro.org wrote:

OK - so we've double checked and re-seated the boards and done some diagnostics. We've found the following:

• On the new board, which contains nodes 8, 9, 10 and 11, only 9 and 10

consistently come up and get served an ip address based on their MAC address. Nodes 8 and 11 don't always show up when trying to probe MAC addresses.

• Nodes 12, 13, 14 and 15 are showing up on a macaddr lookup consistently,

but only sometimes being served an ip address.

• If we probe nodes in the 12-15 range, it sometimes brings them to life.

So, conclusions:

• The new board has only 2 nodes that work consistently
• The board with nodes 12->15 is not behaving.

On a side note, we've noticed that several nodes don't consistently get served, but the ones detailed are the worst offenders.

Not totally sure where to go from here.

Dave

On 5 Jun 2013, at 09:57, Dave Pigott dave.pigott@linaro.org wrote:

Something very odd here, post the broken node replacement. Matt and I are investigating and need exclusive access today, because we need to power cycle it, probably more than once.

Thanks

Dave

On 4 Jun 2013, at 21:45, Matthew Gretton-Dann < matthew.gretton-dann@linaro.org> wrote:

Node 16: 192.168.1.111 Node 17: 192.168.2.126

Hi Dave

thanks for the clarification. Could you get in touch with chris.oakley@calxeda.com directly and describe what is the issue you are experiencing?

Thanks, Ilias

-- Ilias Biris - ilias.biris@linaro.org Technical Program Manager, Linaro M: +358504839608, IRC: ibiris, Skype: ilias_biris Linaro.org | Open source software for ARM SoCs

On 5 June 2013 13:35, Dave Pigott dave.pigott@linaro.org wrote:

No, we just swapped the hardware out. We have a diagram of the node/slot layout as a guide.

Dave

On 5 Jun 2013, at 11:16, Ilias Biris ilias.biris@linaro.org wrote:

https://jira.calxeda.com/browse/LINARO-1

Back to Calxeda and ask their input. For the record, in the same ticket Chris Oakley from Calxeda who's following this case had requested that we arrange a chat with him to coordinate the swap out and walk through FW upgrade after, did that chat take place?

Best regards,

-- Ilias Biris - ilias.biris@linaro.org Technical Program Manager, Linaro M: +358504839608, IRC: ibiris, Skype: ilias_biris Linaro.org | Open source software for ARM SoCs

On 5 June 2013 13:10, Dave Pigott dave.pigott@linaro.org wrote:

...

OK - so we've double checked and re-seated the boards and done some diagnostics. We've found the following:

• On the new board, which contains nodes 8, 9, 10 and 11, only 9 and 10

consistently come up and get served an ip address based on their MAC address. Nodes 8 and 11 don't always show up when trying to probe MAC addresses.

• Nodes 12, 13, 14 and 15 are showing up on a macaddr lookup

consistently, but only sometimes being served an ip address.

• If we probe nodes in the 12-15 range, it sometimes brings them to life.

So, conclusions:

• The new board has only 2 nodes that work consistently
• The board with nodes 12->15 is not behaving.

On a side note, we've noticed that several nodes don't consistently get served, but the ones detailed are the worst offenders.

Not totally sure where to go from here.

Dave

On 5 Jun 2013, at 09:57, Dave Pigott dave.pigott@linaro.org wrote:

Something very odd here, post the broken node replacement. Matt and I are investigating and need exclusive access today, because we need to power cycle it, probably more than once.

Thanks

Dave

On 4 Jun 2013, at 21:45, Matthew Gretton-Dann < matthew.gretton-dann@linaro.org> wrote:

Node 16: 192.168.1.111 Node 17: 192.168.2.126