[202205] dhcrelay: Don't look up the ifindex for the fallback interface (#17797)#18416
Merged
yxieca merged 1 commit intosonic-net:202205from Mar 22, 2024
Merged
Conversation
…-net#17797) Currently, whenever isc-dhcp-relay forwards a packet upstream, internally, it will try to send it on a "fallback" interface. My understanding is that this isn't meant to be a real interface, but instead is basically saying to use Linux's regular routing stack to route the packet appropriately (rather than having isc-dhcp-relay specify specifically which interface to use). The problem is that on systems with a weak CPU, a large number of interfaces, and many upstream servers specified, this can introduce a noticeable delay in packets getting sent. The delay comes from trying to get the ifindex of the fallback interface. In one test case, it got to the point that only 2 packets could be processed per second. Because of this, dhcrelay will easily get backlogged and likely get to a point where packets get dropped in the kernel. Fix this by adding a check saying if we're using the fallback interface, then don't try to get the ifindex of this interface. We're never going to have an interface named this in SONiC. Signed-off-by: Saikrishna Arcot <sarcot@microsoft.com>
saiarcot895
approved these changes
Mar 21, 2024
Contributor
Author
|
@yxieca Could you please help to merge this PR? |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Currently, whenever isc-dhcp-relay forwards a packet upstream, internally, it will try to send it on a "fallback" interface. My understanding is that this isn't meant to be a real interface, but instead is basically saying to use Linux's regular routing stack to route the packet appropriately (rather than having isc-dhcp-relay specify specifically which interface to use).
The problem is that on systems with a weak CPU, a large number of interfaces, and many upstream servers specified, this can introduce a noticeable delay in packets getting sent. The delay comes from trying to get the ifindex of the fallback interface. In one test case, it got to the point that only 2 packets could be processed per second. Because of this, dhcrelay will easily get backlogged and likely get to a point where packets get dropped in the kernel.
Fix this by adding a check saying if we're using the fallback interface, then don't try to get the ifindex of this interface. We're never going to have an interface named this in SONiC.
Why I did it
Manually backport of this PR: #17797
Currently, whenever isc-dhcp-relay forwards a packet upstream, internally, it will try to send it on a "fallback" interface. My understanding is that this isn't meant to be a real interface, but instead is basically saying to use Linux's regular routing stack to route the packet appropriately (rather than having isc-dhcp-relay specify specifically which interface to use).
The problem is that on systems with a weak CPU, a large number of interfaces, and many upstream servers specified, this can introduce a noticeable delay in packets getting sent. The delay comes from trying to get the ifindex of the fallback interface. In one test case, it got to the point that only 2 packets could be processed per second. Because of this, dhcrelay will easily get backlogged and likely get to a point where packets get dropped in the kernel.
Fix this by adding a check saying if we're using the fallback interface, then don't try to get the ifindex of this interface. We're never going to have an interface named this in SONiC.
Work item tracking
How I did it
How to verify it
Which release branch to backport (provide reason below if selected)
Tested branch (Please provide the tested image version)
Description for the changelog
Link to config_db schema for YANG module changes
A picture of a cute animal (not mandatory but encouraged)