Extreme 9k

[mayafox]

Added in Extreme 9920, 8520, and 8720 products, fixed incorrect wattage for 7400-48y, and removed mentions of AC from the model and parts for the MLXe and SLX 9xxx boxes.

[mayafox]

Any guidance on resolving this? last I checked 0.01 would go into 10.11 just fine, and the yaml matches up with the sample output format shown in the main readme.md

weight:
  - value: 10.11
    unit: kg

device-types/Extreme Networks/8520-48Y-8C.yaml failed validation: 10.11 is not a multiple of 0.01

Failed validating 'multipleOf' in schema['properties']['weight']['items']['properties']['value']:
{'minimum': 0, 'multipleOf': 0.01, 'type': 'number'}

On instance['weight'][0]['value']:
10.11

Did a bit of poking around, changing the value to 1011 <<< integer, and not a float, caused the check to pass. This seems to potentially be an issue in jsonschema,

https://github.com/python-jsonschema/jsonschema/blob/main/jsonschema/_validators.py line 181

If I am reading the code right jsonschema is running into a rounding error.

db = 0.01
instance = 10.11
isinstance(db, float)
True
quotient = instance / db
quotient
1010.9999999999999
int(quotient)
1010
int(round(quotient, 2))
1011

[danner26]

Hey @mayafox, the weight definition is brand new and we are clearly still tackling some issues around that. I will use this as a testing ground to fix that issue as well.

I also would just like to confirm, the correct weight should be 10.11 KG, right?

[danner26]

@mayafox I have merged PR #1069 which resolves your issue. We can use multipleOfPrecision instead of multipleOf which resolves the issue. I have tested this with 10, 10.1, 10.11, etc.

[danner26]

@mayafox I started to update some of the weights but I do not know the actual values. I have merged the master branch back into yours, so you should be able to correct the weight values now. Can you please update the weights?

Please also make sure the weights are as accurate as possible, and in a comment on the device, please add a link to the device documentation. Thanks!

[mayafox]

@danner26
Changes in support of fixing the weight look good!
Device documentation should be the hardware information URL, and not the OS admin guides I presume?
Should I call out the different power/airflow combinations as unique models or is AC w/ front-to-rear ok?

Also what's the preferred format for the comments you requested?
comments: '[Extreme Networks X440-G2 Technical Specifications](https://www.extremenetworks.com/product/x440-g2/)'
or
comments: https://www.extremenetworks.com/support/documentation/5520/

[danner26]

Hardware info url yes, but if you have both feel free to throw them both in!

…

On Thu, Jan 26, 2023, 19:00 mayafox ***@***.***> wrote: @danner26 <https://github.com/danner26> Changes in support of fixing the weight look good! Device documentation should be the hardware information URL, and not the OS admin guides I presume? — Reply to this email directly, view it on GitHub <#1067 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB6JNI2WZZQWX7KYRKGRRF3WUMFYZANCNFSM6AAAAAAUG6X5S4> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[cwegener]

@danner26 Just a quick note that you didn't actually "fix" the original issue with the multipleOf keyword. In #1069 you have unknowingly disabled the multipleOf keyword.

There is no such thing as a multipleOfPrecision keyword in the JSON Schema specification.

[danner26]

@cwegener Thanks for the heads up, I am guessing that this issue isnt an actual solution then. Any idea why it works correctly then?

Do you have a proper solution for this?

[cwegener]

Do you have a proper solution for this?

Nope. I just noticed the obvious error and thought I point it out.

I'm in a similar situation like yourself. I'm aware of how floating point arithmetic works in Python (and any other language using binary for floating point) but also have no experience on how to properly apply libraries like jsonschema as part of the whole ecosystem.

My case is even more complex as I ran into this issue in the context of a distributed system where the JSON Schema spec is the 'connecting' tissue but the implementations that send data to each other are blissfully unaware of each other. (https://github.com/singer-io/getting-started/blob/master/docs/SPEC.md)

[danner26]

Gotcha, I appreciate you pointing it out @cwegener ! Here's to hoping we find a solution

[danner26]

@mayafox we fixed the decimal issue, please try applying the proper weight values again. Thank you for your patience

Also @cwegener feel free to take a look at our implementation, one of our devs was able to reverse engineer and fix the issue for us.

[mayafox]

@danner26 Done!

[danner26]

I am a little confused, what is going on with all of the renames such as the BR-MLXE-16-MR2-AC being renamed to BR-MLXE-16-MR2? I see the BR-MLXE-16-MR2-AC listed here.

We shouldnt be removing or renaming device definitions, you should add new files for different devices. Unless I am missing something. Can you explain? @mayafox

[mayafox]

So there's a couple issues here with the -AC models. I chatted with one of my colleagues on adding in all the various Power supply (AC/DC) and air-flow combinations that are listed, and they viewed it as a moot point and not worth the effort (not sure they fully understood the question either). I also asked you what the projects preference was, but you missed that question as well. I don't mind doing the work to add in the variations, but a clear direction would be good. For the moment I dropped the -AC from the previously submitted items, and will await your response on what I need to do on this matter.

[danner26]

Sorry about that, I totally missed that questions. So the only thing the -AC is used for is when the AC power module is installed? If that is the case, then I agree it is a moot point. I think modelling them without the AC makes sense, and then the modules that you install (AC or DC power supply) is what will determine which power type it uses. Is that understanding where your changes came from?

[mayafox]

Yes, and for some boxes (MLX and the 9920) it is possible to have both AC and DC power supplies installed and in service. Airflow is generally uniform, and can impact the power supplies choices as well, but most of the platforms allow you to switch by swapping the AC PSU for DC, and swapping fans and PSU's to change the air flow. It's a mess of permutations I would rather not poke if I can avoid it.
For information purposes I am modeling them with the correct wattages, and AC connections and front to rear airflow, and max possible weight.

[danner26]

So I see what the mess is here, this is quite the issue. Let me ask you this, if I were to go to the vendor and ask for a BR-MLXE-16-MR2-AC, would they provide me with the unit and the AC power supplies? Like do they actually list for sale BR-MLXE-16-MR2-AC that includes all of the kit, plus the BR-MLXE-16-MR2 which doesnt include the PSU? Or do you have to buy either a BR-MLXE-16-MR2-AC or a BR-MLXE-16-MR2-DC?

Because this library acts as a source of truth for a specific device, I do not think we should be removing/renaming legitimate models even if they do have the option to be changed on the fly. If a product can be purchased by default under the BR-MLXE-16-MR2-AC sku, it should have its own file. Just as if you can buy the product without the AC/DC PSUs, you should also have another file that is BR-MLXE-16-MR2.

Does that make sense? Unless you cannot buy a BR-MLXE-16-MR2-AC and you can only buy a BR-MLXE-16-MR2, then I think ultimately you should revert the renames/changes in those files.

[mayafox]

@danner26
The TLDR answer is yeap it makes sense and I will update the platform files accordingly. Thanks for providing a clear direction on what's expected. Might take me a few days to get the work done though, the cube farm is demanding more of my time.

"Let me ask you this, if I were to go to the vendor and ask for a BR-MLXE-16-MR2-AC, would they provide me with the unit and the AC power supplies?" Based of the ordering information page from the data sheet, yeap, they used to list the bare chassis sans mgmt, and power, but that's been removed,

BR-MLXE-16-MR2-M-AC
MLXe-16 AC system with one MR2 (M) management module, three high-speed switch fabric modules, four 1,800W
AC power supplies, two exhaust fan assembly kits, and air filter. Power cord not included.
BR-MLXE-16-MR2-M-DC
MLXe-16 DC system with one MR2 (M) management module, three high-speed switch fabric modules, four 1,800W
DC power supplies, two exhaust fan assembly kits, and air filter. Power cord not included.
BR-MLXE-16-MR2-X-AC
MLXe-16 AC system with one MR2 (X) management module, three high-speed switch fabric modules, four 1,800W
AC power supplies, two exhaust fan assembly kits, and air filter. Power cord not included.
BR-MLXE-16-MR2-X-DC
MLXe-16 DC system with one MR2 (X) management module, three high-speed switch fabric modules, four 1,800W
DC power supplies, two exhaust fan assembly kits, and air filter. Power cord not included.

[mayafox]

@danner26
Does this fall in line with what you expect? If so, I'll work on the rest of the Extreme boxes.

[danner26]

Hey @mayafox take your time.. let me know when you are ready for me to review it again
And yes that looks like what I am thinking, whatever is the default configurations from the vendor is what we should define. The idea is that this is a repo that you can import the standardized models based off of, and then after imported and created in NetBox, the user can modify the device in NetBox from the default configuration to model the real world. Let me know if you need any clarification, I hope I am making sense

[github-actions]

This PR has been automatically marked as stale because it has not had recent activity. It will be closed automatically if no further progress is made.

[danner26]

Hello, unfortunately today we had to make a change to the schema in which weight is defined by. The upstream NetBox API requires weight to be passed to it in the following format:

weight: 12.01
weight_unit: lb

Due to this change, I have updated your branch against the current master. This might require you to update your PR, if you had weight currently defined, so that your PR is in-line with the new weight schema. If you have any issues, please let me know or refer to this PR for more information.

[mayafox]

Closing for rework of the code.