Questions about openbmc-test-automation

Questions about openbmc-test-automation

[Tony Lee (李文富)]

[-- Attachment #1: Type: text/plain, Size: 564 bytes --]

Hi Rahul,

I'm wondering that why "REDFISH_NW_ETH0" is hardcode as "Managers/bmc/EthernetInterfaces/eth0/" in /data/variables.py
https://github.com/openbmc/openbmc-test-automation/blob/master/data/variables.py#L155
Furthermore, the command "lan print" didn't specify the channel number in /ipmi/test_ipmi_network.robot.
Therefore, the default will print information on the first found LAN channel.

Isn't it more reasonable to set the interface and the channel number as environment variables and give them default values?

Thanks
Best Regards,
Tony


[-- Attachment #2: Type: text/html, Size: 3349 bytes --]

Re: Questions about openbmc-test-automation

[Rahul Maheshwari]

[-- Attachment #1: Type: text/plain, Size: 1151 bytes --]

Thanks for pointing that out. Passing interface/channel number is an easier
option but we should be avoiding passing environment variable unless there
is no other way. I would say better way to overcome this is to figure out
the channel count using "ip addr" command from BMC cli.
Let us know if that sounds good to you.
Also let us know if you would like to drop code change for the same?

Thanks
Rahul






On Tue, Nov 5, 2019 at 2:28 PM Tony Lee (李文富) <Tony.Lee@quantatw.com> wrote:

> Hi Rahul,
>
>
>
> I'm wondering that why "REDFISH_NW_ETH0" is hardcode as
> "Managers/bmc/EthernetInterfaces/eth0/" in /data/variables.py
>
>
> https://github.com/openbmc/openbmc-test-automation/blob/master/data/variables.py#L155
>
> Furthermore, the command "lan print" didn't specify the channel number in
> /ipmi/test_ipmi_network.robot.
>
> Therefore, the default will print information on the first found LAN
> channel.
>
>
>
> Isn't it more reasonable to set the interface and the channel number as
> environment variables and give them default values?
>
>
>
> Thanks
>
> Best Regards,
>
> Tony
>
>
>

[-- Attachment #2: Type: text/html, Size: 2672 bytes --]

RE: Questions about openbmc-test-automation

[Tony Lee (李文富)]

Hi Rahul,

For example, case "Retrieve IP Address Via IPMI And Verify Using Redfish" in /ipmi/test_ipmi_network.robot.
The new way is as follows:

IPMI network channel logically starts from 1. For example, once it figure out the channel count is 3,
we can retrieve IP addresses through channels 1, 2, and 3, and verify them correspond to eth0, eth1, and eth2, respectively by redfish.
Is it right?

Thanks
Best Regards,
Tony

From: Rahul Maheshwari <rahulmaheshwari01@gmail.com> 
Sent: Tuesday, November 5, 2019 6:51 PM
To: Tony Lee (李文富) <Tony.Lee@quantatw.com>
Cc: openbmc@lists.ozlabs.org
Subject: Re: Questions about openbmc-test-automation

Thanks for pointing that out. Passing interface/channel number is an easier option but we should be avoiding passing environment variable unless there is no other way. I would say better way to overcome this is to figure out the channel count using "ip addr" command from BMC cli. 
Let us know if that sounds good to you.
Also let us know if you would like to drop code change for the same? 

Thanks
Rahul




 

On Tue, Nov 5, 2019 at 2:28 PM Tony Lee (李文富) <mailto:Tony.Lee@quantatw.com> wrote:
Hi Rahul,
 
I'm wondering that why "REDFISH_NW_ETH0" is hardcode as "Managers/bmc/EthernetInterfaces/eth0/" in /data/variables.py
https://github.com/openbmc/openbmc-test-automation/blob/master/data/variables.py#L155
Furthermore, the command "lan print" didn't specify the channel number in /ipmi/test_ipmi_network.robot.
Therefore, the default will print information on the first found LAN channel. 
 
Isn't it more reasonable to set the interface and the channel number as environment variables and give them default values?
 
Thanks
Best Regards,
Tony
 

Re: Questions about openbmc-test-automation

[Rahul Maheshwari]

[-- Attachment #1: Type: text/plain, Size: 1978 bytes --]

Yes, that is correct way.

On Wed, Nov 6, 2019 at 11:22 AM Tony Lee (李文富) <Tony.Lee@quantatw.com>
wrote:

> Hi Rahul,
>
> For example, case "Retrieve IP Address Via IPMI And Verify Using Redfish"
> in /ipmi/test_ipmi_network.robot.
> The new way is as follows:
>
> IPMI network channel logically starts from 1. For example, once it figure
> out the channel count is 3,
> we can retrieve IP addresses through channels 1, 2, and 3, and verify them
> correspond to eth0, eth1, and eth2, respectively by redfish.
> Is it right?
>
> Thanks
> Best Regards,
> Tony
>
> From: Rahul Maheshwari <rahulmaheshwari01@gmail.com>
> Sent: Tuesday, November 5, 2019 6:51 PM
> To: Tony Lee (李文富) <Tony.Lee@quantatw.com>
> Cc: openbmc@lists.ozlabs.org
> Subject: Re: Questions about openbmc-test-automation
>
> Thanks for pointing that out. Passing interface/channel number is an
> easier option but we should be avoiding passing environment variable unless
> there is no other way. I would say better way to overcome this is to figure
> out the channel count using "ip addr" command from BMC cli.
> Let us know if that sounds good to you.
> Also let us know if you would like to drop code change for the same?
>
> Thanks
> Rahul
>
>
>
>
>
>
> On Tue, Nov 5, 2019 at 2:28 PM Tony Lee (李文富) <mailto:
> Tony.Lee@quantatw.com> wrote:
> Hi Rahul,
>
> I'm wondering that why "REDFISH_NW_ETH0" is hardcode as
> "Managers/bmc/EthernetInterfaces/eth0/" in /data/variables.py
>
> https://github.com/openbmc/openbmc-test-automation/blob/master/data/variables.py#L155
> Furthermore, the command "lan print" didn't specify the channel number in
> /ipmi/test_ipmi_network.robot.
> Therefore, the default will print information on the first found LAN
> channel.
>
> Isn't it more reasonable to set the interface and the channel number as
> environment variables and give them default values?
>
> Thanks
> Best Regards,
> Tony
>
>

[-- Attachment #2: Type: text/html, Size: 2800 bytes --]

RE: Questions about openbmc-test-automation

[Tony Lee (李文富)]

Hi Rahul,

It sounds good to me and I would like to drop code change.

From: Rahul Maheshwari <rahulmaheshwari01@gmail.com> 
Sent: Wednesday, November 6, 2019 3:13 PM
To: Tony Lee (李文富) <Tony.Lee@quantatw.com>
Cc: openbmc@lists.ozlabs.org
Subject: Re: Questions about openbmc-test-automation

Yes, that is correct way.

On Wed, Nov 6, 2019 at 11:22 AM Tony Lee (李文富) <mailto:Tony.Lee@quantatw.com> wrote:
Hi Rahul,

For example, case "Retrieve IP Address Via IPMI And Verify Using Redfish" in /ipmi/test_ipmi_network.robot.
The new way is as follows:

IPMI network channel logically starts from 1. For example, once it figure out the channel count is 3,
we can retrieve IP addresses through channels 1, 2, and 3, and verify them correspond to eth0, eth1, and eth2, respectively by redfish.
Is it right?

Thanks
Best Regards,
Tony

From: Rahul Maheshwari <mailto:rahulmaheshwari01@gmail.com> 
Sent: Tuesday, November 5, 2019 6:51 PM
To: Tony Lee (李文富) <mailto:Tony.Lee@quantatw.com>
Cc: mailto:openbmc@lists.ozlabs.org
Subject: Re: Questions about openbmc-test-automation

Thanks for pointing that out. Passing interface/channel number is an easier option but we should be avoiding passing environment variable unless there is no other way. I would say better way to overcome this is to figure out the channel count using "ip addr" command from BMC cli. 
Let us know if that sounds good to you.
Also let us know if you would like to drop code change for the same? 

Thanks
Rahul




 

On Tue, Nov 5, 2019 at 2:28 PM Tony Lee (李文富) <mailto:mailto:Tony.Lee@quantatw.com> wrote:
Hi Rahul,
 
I'm wondering that why "REDFISH_NW_ETH0" is hardcode as "Managers/bmc/EthernetInterfaces/eth0/" in /data/variables.py
https://github.com/openbmc/openbmc-test-automation/blob/master/data/variables.py#L155
Furthermore, the command "lan print" didn't specify the channel number in /ipmi/test_ipmi_network.robot.
Therefore, the default will print information on the first found LAN channel. 
 
Isn't it more reasonable to set the interface and the channel number as environment variables and give them default values?
 
Thanks
Best Regards,
Tony
 

Re: Questions about openbmc-test-automation

[Rahul Maheshwari]

[-- Attachment #1: Type: text/plain, Size: 2472 bytes --]

Sure. Thanks for taking up the initiative.

On Thu, 7 Nov 2019, 07:07 Tony Lee (李文富), <Tony.Lee@quantatw.com> wrote:

> Hi Rahul,
>
> It sounds good to me and I would like to drop code change.
>
> From: Rahul Maheshwari <rahulmaheshwari01@gmail.com>
> Sent: Wednesday, November 6, 2019 3:13 PM
> To: Tony Lee (李文富) <Tony.Lee@quantatw.com>
> Cc: openbmc@lists.ozlabs.org
> Subject: Re: Questions about openbmc-test-automation
>
> Yes, that is correct way.
>
> On Wed, Nov 6, 2019 at 11:22 AM Tony Lee (李文富) <mailto:
> Tony.Lee@quantatw.com> wrote:
> Hi Rahul,
>
> For example, case "Retrieve IP Address Via IPMI And Verify Using Redfish"
> in /ipmi/test_ipmi_network.robot.
> The new way is as follows:
>
> IPMI network channel logically starts from 1. For example, once it figure
> out the channel count is 3,
> we can retrieve IP addresses through channels 1, 2, and 3, and verify them
> correspond to eth0, eth1, and eth2, respectively by redfish.
> Is it right?
>
> Thanks
> Best Regards,
> Tony
>
> From: Rahul Maheshwari <mailto:rahulmaheshwari01@gmail.com>
> Sent: Tuesday, November 5, 2019 6:51 PM
> To: Tony Lee (李文富) <mailto:Tony.Lee@quantatw.com>
> Cc: mailto:openbmc@lists.ozlabs.org
> Subject: Re: Questions about openbmc-test-automation
>
> Thanks for pointing that out. Passing interface/channel number is an
> easier option but we should be avoiding passing environment variable unless
> there is no other way. I would say better way to overcome this is to figure
> out the channel count using "ip addr" command from BMC cli.
> Let us know if that sounds good to you.
> Also let us know if you would like to drop code change for the same?
>
> Thanks
> Rahul
>
>
>
>
>
>
> On Tue, Nov 5, 2019 at 2:28 PM Tony Lee (李文富) <mailto:mailto:
> Tony.Lee@quantatw.com> wrote:
> Hi Rahul,
>
> I'm wondering that why "REDFISH_NW_ETH0" is hardcode as
> "Managers/bmc/EthernetInterfaces/eth0/" in /data/variables.py
>
> https://github.com/openbmc/openbmc-test-automation/blob/master/data/variables.py#L155
> Furthermore, the command "lan print" didn't specify the channel number in
> /ipmi/test_ipmi_network.robot.
> Therefore, the default will print information on the first found LAN
> channel.
>
> Isn't it more reasonable to set the interface and the channel number as
> environment variables and give them default values?
>
> Thanks
> Best Regards,
> Tony
>
>

[-- Attachment #2: Type: text/html, Size: 3771 bytes --]