Move to json storage, address pblindex issue #721
Conversation
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
sknat
force-pushed
the
nsk-move-json-storage
branch
from
6a4c197 to
3fe1285
Compare
sknat
force-pushed
the
nsk-move-json-storage
branch
from
3fe1285 to
0c31407
Compare
sknat
force-pushed
the
nsk-move-json-storage
branch
4 times, most recently
from
579cb63 to
b842c3d
Compare
sknat
force-pushed
the
nsk-move-json-storage
branch
3 times, most recently
from
4adc912 to
5d8435c
Compare
sknat
force-pushed
the
nsk-move-json-storage
branch
from
5d8435c to
b400da8
Compare
This patch changes the way we persist the data on disk when running Calico/VPP. Instead of using struc and binary format we transition to json files. Size should not be an issue as number of pods per node are typically low (~100). This will make troubleshooting easier and errors clearer when parsing fails. We thus remove the /bin/debug troubleshooting utility as the data format is not human readable. Doing this, we address an issue where PBL indexes were reused upon dataplane restart, as they were stored in a list. We now will use a map to retain the containerIP mapping. We also split the configuration from runtime spec in LocalPodSpec and add a step to clear it when corresponding VRFs are not found in VPP. Finally we address an issue where uRPF was not properly set up for ipv6. Signed-off-by: Nathan Skrzypczak <[email protected]>
sknat
force-pushed
the
nsk-move-json-storage
branch
from
b400da8 to
f4b7b9b
Compare
hedibouattour
requested changes
Sep 17, 2025
| @@ -39,7 +39,7 @@ func (v *VppLink) SetCustomURPF(swifindex uint32, tableID uint32) error { | |||
| return nil | |||
| } | |||
|
|
|||
| func (v *VppLink) UnsetURPF(swifindex uint32) error { | |||
| func (v *VppLink) UnsetURPF(swifindex uint32, ipFamily IPFamily) error { | |||
There was a problem hiding this comment.
Good catch, but I think you forgot to consume this one.
| } | ||
| // We do not have a VRF in VPP for this pod, VPP has probably | ||
| // restarted, so we clear the state we have. | ||
| podSpec.LocalPodSpecStatus = *model.NewLocalPodSpecStatus() |
There was a problem hiding this comment.
Why assume that vpp restarted in this case? That's the normal path for a freshly created pod, no ?
|
Amazing change, thanks Nathan ! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
This patch changes the way we persist the data on disk when running Calico/VPP. Instead of using struc and binary format we transition to json files. Size should not be an issue as number of pods per node are typically low (~100). This will make troubleshooting easier and errors clearer when parsing fails.
We thus remove the /bin/debug troubleshooting utility as the data format is not human readable.
Doing this, we address an issue where PBL indexes were reused upon dataplane restart, as they were stored in a list. We now will use a map to retain the containerIP mapping.
We also split the configuration from runtime spec in LocalPodSpec and add a step to clear it when corresponding VRFs are not found in VPP.
Finally we address an issue where uRPF was not properly set up for ipv6.