Clean git merge#8
Closed
jluhrsen wants to merge 238 commits into
Closed
Conversation
The document is refactored to meet the ovn-org feature template [0]. [0] https://github.com/ovn-org/ovn-kubernetes/blob/master/docs/features/template.md Signed-off-by: Ram Lavi <ralavi@redhat.com>
Signed-off-by: Ram Lavi <ralavi@redhat.com>
This picks up the following relevant bug fixes: https://issues.redhat.com/browse/FDP-906 "ovn-controller: lib/ovsdb-idl.c:3596: assertion row->new_datum != NULL failed in ovsdb_idl_txn_write__()" 6448f5e364 pinctrl: Skip non-local mac bindings in run_buffered_binding(). ea35347320 pinctrl: Skip deleted mac bindings in run_buffered_binding(). 33a6ae53f4 pinctrl: Use correct map size in pinctrl_handle_put_fdb(). 8eaa7d5991 controller: Fix "use after free" issue in statctrl_run(). 8579859f51 mac-cache: Properly handle deletion of SB mac_bindings. https://issues.redhat.com/browse/FDP-752 "ovn-northd IPAM incorrectly reports duplicate IP when part of excluded_ips" 2a24b03f7f ipam: Do not report error for static assigned IPs. https://issues.redhat.com/browse/FDP-786 "When an ECMP symmetric route is removed, northd removes all logical flows from SBDB for ECMP" 7b00627433 northd: Respect --ecmp-symmetric-reply for single routes. Signed-off-by: Dumitru Ceara <dceara@redhat.com>
Signed-off-by: arkadeepsen <arsen@redhat.com>
OCPBUGS-42707: Bump OVN to ovn24.09-24.09.1-10.el9fdp
Signed-off-by: Flavio Fernandes <ffernandes@nvidia.com>
Signed-off-by: Flavio Fernandes <ffernandes@nvidia.com>
This reverts commit 1243011.
OCPBUGS-48330,OCPBUGS-42609,OCPBUGS-46585,SDN-4930: Downstream Merge [01-23-2025]
Fixes a null pointer exception when network policy port has no protocol. If the protocol is missing in the network policy port definition, it should be assumed to be TCP. Signed-off-by: Tim Rozet <trozet@redhat.com>
ShallowClone has to copy all factories. Signed-off-by: Patryk Diak <pdiak@redhat.com>
Commit 6dda0b5 ("factory: Bump the event queue size to 1K.") increased the event queue size to 1K events. However, in combination with fe17136 ("factory: Reduce contention on informer locks.") which configures 201 internal informers this might end up using too much memory in cases when controllers cannot consume events as fast as they're queued by the kube API. For each kubernetes API object type we consume: N_internal_informers x N_queues x N_events x sizeof(event) memory. That currently translates to: N_internal_informers = 201 N_queues = 15 N_events = 1000 sizeof(event) = 32B => ~92MB of memory per object type Given that ovn-kubernetes processes need to be informed about multiple object types this can grow to a significantly large number when controllers that are supposed to consume events from the internal informer queues are slow. Reduce the queue size, making it 100, in order to lower the worst case scenario memory usage: N_internal_informers = 201 N_queues = 15 N_events = 100 sizeof(event) = 32B => ~9.2MB of memory per object type Signed-off-by: Dumitru Ceara <dceara@redhat.com>
Clone raFactory in ShallowClone
Signed-off-by: Patryk Diak <pdiak@redhat.com>
Clone frrFactory in ShallowClone
factory: Set default event queue size to 100.
Previously, if a new NAD was added to an existing network after a pod referencing it, the pod would never start. This is fixed by reconciling pending pods when the secondary network controller reconciles a new NAD. Signed-off-by: Patryk Diak <pdiak@redhat.com>
Fix doc: Replace ovn-org with ovn-kubernetes to reflect repo move
Reconcile pending pods when a NAD is added to an existing network
Fixes NPE seen at: openshift#2427 (comment) Certain network types may not have a pod handler or retry framework for cluster manager. Signed-off-by: Tim Rozet <trozet@redhat.com>
Fixes NPE seen at: openshift#2427 (comment) Certain network types may not have a pod handler or retry framework for cluster manager. Signed-off-by: Tim Rozet <trozet@redhat.com>
SDN-4930: Downstream Merge [01-28-2025]
Compare annotations directly if possible. For network specific map entries only compare raw json entries without parsing the map in full. Co-authored-by: Tim Rozet <trozet@redhat.com> Signed-off-by: Patryk Diak <pdiak@redhat.com>
Instead of always parsing all node/join subnets parse the raw json map and only compute the results for the affected network. Signed-off-by: Patryk Diak <pdiak@redhat.com>
Signed-off-by: Patryk Diak <pdiak@redhat.com>
udn: Add event when ip range full
endPort was already supported by the NetworkPolicy, all we need to do is copy the new field for the netpol handler. Signed-off-by: Nadia Pinaeva <npinaeva@redhat.com>
…anup dist/images: fix cleanup command for northd
Using convention we already use elsewhere, rename PreStart to Init and call from already existing controller manager init method. Signed-off-by: Jaime Caamaño Ruiz <jcaamano@redhat.com>
Signed-off-by: Jaime Caamaño Ruiz <jcaamano@redhat.com>
Signed-off-by: Jaime Caamaño Ruiz <jcaamano@redhat.com>
As is not sensitive or have dependencies with any other thing Signed-off-by: Jaime Caamaño Ruiz <jcaamano@redhat.com>
These are types that are used by both the gateway and the management port code. Export them as the plan is to have management port on its own package. Signed-off-by: Jaime Caamaño Ruiz <jcaamano@redhat.com>
There is a bug in ParseDump that fails to parse if interface names are quoted. We want to use ParseDump on our tests. Since our interface names specifically don't need to be quoted, remove the quotes. Signed-off-by: Jaime Caamaño Ruiz <jcaamano@redhat.com>
These tests need as precondition the management port nft entities. Use knftables.ParseDump instead of setting up a fully fledged management port. Signed-off-by: Jaime Caamaño Ruiz <jcaamano@redhat.com>
Add the new management port package interface and use it externally. Signed-off-by: Jaime Caamaño Ruiz <jcaamano@redhat.com>
Move the existing implementaiton to a new package. It should be functionally equivalent except: * nft sets are configured as early as possible, the rest is configured upon Start (synchronously) * moved management port routing table, rules, routes etc from gateway code to this package, configued upon Start (synchronously) Signed-off-by: Jaime Caamaño Ruiz <jcaamano@redhat.com>
Signed-off-by: Jaime Caamaño Ruiz <jcaamano@redhat.com>
Multinetpol: update to the latest API version that adds endPort
Signed-off-by: Surya Seetharaman <suryaseetharaman.9@gmail.com>
This contains the following relevant changes:
- controller: Fix active mac-binding refresh for IPv6.
- controller: Send ARP/ND for stale mac_bindings entries. (#FDP-1135)
https://issues.redhat.com/browse/FDP-1135
- northd: Fix action parsing in build_lb_vip_actions(). (#FDP-1095)
https://issues.redhat.com/browse/FDP-1095
- northd: Fix missing tier related ACL flows. (#FDP-1154)
https://issues.redhat.com/browse/FDP-1154
- mac-cache: Fix expiration of active FDB entry due to skipped update. (#FDP-1132)
https://issues.redhat.com/browse/FDP-1132
- mac-cache: Fix expiration of active MAC binding due to skipped update. (#FDP-1130)
https://issues.redhat.com/browse/FDP-1130
- mac-cache: Fix MAC binding entry lookup for timestamp refresh. (#FDP-1131)
https://issues.redhat.com/browse/FDP-1131
Signed-off-by: Dumitru Ceara <dceara@redhat.com>
Add BGP Tests: Phase2
OCPBUGS-48678: Update OVN to FDP25.A.1 24.09.2-41
OCPBUGS-54245, SDN-5772: Downstream merge 2025-03-21
…openshift-4.19-ose-ovn-kubernetes OCPBUGS-45741: Updating ose-ovn-kubernetes-container image to be consistent with ART for 4.19
Full sync with 4.19
Owner
Author
|
looks like this is the route we'll take |
jluhrsen
pushed a commit
that referenced
this pull request
May 9, 2025
use NetworkSelector spec to match net-attach-defs instead of explicitly specify a list of net-attach-defs. Signed-off-by: Xiaobin Qu <xqu@nvidia.com> Signed-off-by: Flavio Fernandes <ffernandes@nvidia.com>
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
20 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.
this is the POC showing that once the git merge with all the dups get's merged the next one will be simple and
just keep us in sync. I created a dummy (empty) commit in my local release-4.19 repo and then git
git mergein mylocal 4.18 branch that has the git merge with dups PR already in.
looks clean.