…vice into cache file.

2. resume running service with snapshot cache file by launch params 'init-mode=local'
3. new API(/v2/vs/${SERVICE-ID}/health) for dpvs-healthcheck

Bumps [golang.org/x/net](https://github.com/golang/net) from 0.10.0 to 0.17.0.
- [Commits](golang/net@v0.10.0...v0.17.0)

---
updated-dependencies:
- dependency-name: golang.org/x/net
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <[email protected]>

dpvs-agent dump/launch  services

…gent/golang.org/x/net-0.17.0

build(deps): bump golang.org/x/net from 0.10.0 to 0.17.0 in /tools/dpvs-agent

…ched wrong vs-version from client

…not cached

…ged adminstratively

When healthcheck program updating the rs's weight, it should consider:

* rs's weight may change adminstratively
* rs availabilty may change due to the server is closed or crashed

Healthcheck stores the original weight when it sends down notification in order to restore
the rs's weight when it gets healthy later. The problem is original weight is syncd from
dpvs periodically and not always up-to-date, which may result in incorrect weight update in
healthcheck up notification.

Signed-off-by: ywc689 <[email protected]>

Unfortunately, the CAS solution in the last commit is not effective for healthcheck
process cannot maintain the correct previous weight value. Thus resource version is
used instead, which should be passed to the lb api when the healthcheck try to update
rs, and the api would reject the request if the resouce version is outdated.

use vs version to resync target status when notification fails

Signed-off-by: ywc689 <[email protected]>

- update config version finally in updateConfig and only when the backend config is in healthy state
- fix notification resync fail problem caused by insufficient channel size
- fix array index problem in function NewTargetFromStr

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

…g. POST|PUT /vs/${ID}/rs, PUT /vs/${ID})

Healthcheck api /vs/${ID}/rs/health would not update local cache.

GET API /vs | /vs/${ID} response dpvs-agent local cache info default
and response dpvs running service detail with query param `healthcheck=true`

Signed-off-by: ywc689 <[email protected]>

… services

…checkers

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

…p script will not be excuted

when all old rs(healthcheck is alive) are removed and new rs(healthcheck is alive) add in same reload.

Fix backend update racing problem with dpvs-agent and healthcheck.

use better random source

Signed-off-by: ywc689 <[email protected]>

…ning with gcc version 8.0+

Signed-off-by: ywc689 <[email protected]>

Reduce atomic op by one.

…random with rte_rand

Signed-off-by: ywc689 <[email protected]>

bugfix: new_vs quorum_up script will not be excuted

ip_vs_conn: A small improvement

Signed-off-by: ywc689 <[email protected]>

ipvs: sctp implementation

Bugfixes from Issues

…migration

Signed-off-by: ywc689 <[email protected]>

Ietf quic

tools/ipvsadm: fix could not remove lip

This is a temporary fix, and hopefully can work until Spring 2025.
https://github.blog/changelog/2024-03-07-github-actions-all-actions-will-run-on-node20-instead-of-node16-by-default/

Signed-off-by: ywc689 <[email protected]>

ci: fix unsupported Node20 problem

It fixes following problems in Previous implementation.
1. Neglected the fact that multicast IPv4/IPv6 address can be mapped to one multicast hw address,
   and a lower dpvs port may have multiple upper ports (such as vlan).
2. Multicast hw addresses could sync from kni multiple times or be deleted by mistake.
3. Interferences of linked down kni devices.

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

fix compile error on RHEL 9.4

README:  add ipvsadm RS port

Multicast address bugfix and improvements

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

lldp: add supports for lldp protocol

…ularity

Signed-off-by: ywc689 <[email protected]>

…nularity

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

This problem is caused by underflow of unsigned integer.

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

define buffer size by INET6_ADDRSTRLEN

ipset supports for ipvs acl and dpvs-agent

…ber is over DPVS_MAX_LCORE.

Fixed issue #991.

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

…ation

- configure the addresses only when ipv6 is enabled
- configure the addresses on vlan devices
- configure the addresses on ip-tunnel devices (though nothing is done)

Signed-off-by: ywc689 <[email protected]>

Referencing linux kernel(https://www.kernel.org/doc/Documentation/networking/ip-sysctl.txt),
dpvs supports 4 addr_gen_mode
  - eui64
  - none
  - stable-privacy
  - random

Signed-off-by: ywc689 <[email protected]>

Bugfix 20240902

…sion of dpdk build script

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

V1.9.8

Signed-off-by: ywc689 <[email protected]>

Signed-off-by: ywc689 <[email protected]>

Bugfix 20240919