TE-9.1 : FIB FAILURE DUE TO HARDWARE RESOURCE EXHAUST

[cprabha]

Hi,

This PR is Initial Commit for TE-9.1 FIB FAILURE DUE TO HARDWARE RESOURCE EXHAUST.
Resolves : #508

Thanks,
Prabha

[OpenConfigBot]

Pull Request Functional Test Report for #2048 / 555953e

Virtual Devices

Device	Test	Test Documentation	Job	Raw Log
Arista cEOS		TE-9.1: FIB FAILURE DUE TO HARDWARE RESOURCE EXHAUST
Cisco 8000E		TE-9.1: FIB FAILURE DUE TO HARDWARE RESOURCE EXHAUST
Cisco XRd		TE-9.1: FIB FAILURE DUE TO HARDWARE RESOURCE EXHAUST
Juniper cPTX		TE-9.1: FIB FAILURE DUE TO HARDWARE RESOURCE EXHAUST
Juniper ncPTX		TE-9.1: FIB FAILURE DUE TO HARDWARE RESOURCE EXHAUST
Nokia SR Linux		TE-9.1: FIB FAILURE DUE TO HARDWARE RESOURCE EXHAUST
Openconfig Lemming		TE-9.1: FIB FAILURE DUE TO HARDWARE RESOURCE EXHAUST

Hardware Devices

Device	Test	Test Documentation	Raw Log
Arista 7808		TE-9.1: FIB FAILURE DUE TO HARDWARE RESOURCE EXHAUST
Cisco 8808		TE-9.1: FIB FAILURE DUE TO HARDWARE RESOURCE EXHAUST
Juniper PTX10008		TE-9.1: FIB FAILURE DUE TO HARDWARE RESOURCE EXHAUST
Nokia 7250 IXR-10e		TE-9.1: FIB FAILURE DUE TO HARDWARE RESOURCE EXHAUST

Help

[cprabha]

@greg-dennis This test needs to inject 2.5M routes . As per current guidelines in contributing.md none of address blocks can be used to inject 2.5M routes. I have used 20.0.1.1/32. With this static check fails. please could you suggest how to resolve this.

[coveralls]

Pull Request Test Coverage Report for Build 6175565869

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 45.707%

---

Totals
Change from base Build 6175447171:	0.0%
Covered Lines:	1347
Relevant Lines:	2947

---

💛 - Coveralls

[cprabha]

@greg-dennis This test needs to inject 2.5M routes . As per current guidelines in contributing.md none of address blocks can be used to inject 2.5M routes. I have used 20.0.1.1/32. With this static check fails. please could you suggest how to resolve this.

@greg-dennis , please could you check and help with one more network to use in script to generate 2.5M routes . As of now IP address blocks which are provided in CONTRIBUTING.md does not support 2.5M scale.
There is another test we need similar scale and that needs 4 such IP address groups. . Thank you.

[xw-g]

@greg-dennis This test needs to inject 2.5M routes . As per current guidelines in contributing.md none of address blocks can be used to inject 2.5M routes. I have used 20.0.1.1/32. With this static check fails. please could you suggest how to resolve this.

@greg-dennis , please could you check and help with one more network to use in script to generate 2.5M routes . As of now IP address blocks which are provided in CONTRIBUTING.md does not support 2.5M scale. There is another test we need similar scale and that needs 4 such IP address groups. . Thank you.

@dplore @liulk Please let us know your thoughts on this restriction

@cprabha how about we use BGP to advertise v6 prefix to use up most of the FIB resourcing, and then injecting gRIBI routes? I think that maybe required much less gRIBI v4 routes. What do you think?

@xw-g Test is already validated with v4 bgp routes and , If we need to code with v6 routes will have to re validate test, which will delay this PR and further more we have not tested FIB behavior with v6 . If any issues observed will cause further delay.

If we get new address block in v4 will help to expediate this PR.

Also couple of other BGP scale tests we have test steps with similar scale and new address block is helpful.

@greg-dennis is it convenient to get one more v4 block? If not, @cprabha I think probably you can do v6 over v4 advertisement as a way to fill up the routing table. Thanks.

[cprabha]

@greg-dennis This test needs to inject 2.5M routes . As per current guidelines in contributing.md none of address blocks can be used to inject 2.5M routes. I have used 20.0.1.1/32. With this static check fails. please could you suggest how to resolve this.

@greg-dennis , please could you check and help with one more network to use in script to generate 2.5M routes . As of now IP address blocks which are provided in CONTRIBUTING.md does not support 2.5M scale. There is another test we need similar scale and that needs 4 such IP address groups. . Thank you.

@dplore @liulk Please let us know your thoughts on this restriction

@cprabha how about we use BGP to advertise v6 prefix to use up most of the FIB resourcing, and then injecting gRIBI routes? I think that maybe required much less gRIBI v4 routes. What do you think?

@xw-g Test is already validated with v4 bgp routes and , If we need to code with v6 routes will have to re validate test, which will delay this PR and further more we have not tested FIB behavior with v6 . If any issues observed will cause further delay.
If we get new address block in v4 will help to expediate this PR.
Also couple of other BGP scale tests we have test steps with similar scale and new address block is helpful.

@greg-dennis is it convenient to get one more v4 block? If not, @cprabha I think probably you can do v6 over v4 advertisement as a way to fill up the routing table. Thanks.

@greg-dennis , @xw-g I have updated script with v6 block and all checks passed. Thanks.

[greg-dennis]

Sorry, one nit

[greg-dennis]

Approving for Go quality