Graceful restart is a key event for bgpd, related log print is debug level. To change it to info level to get more visibilities when this kind of event is triggered.
To pick the below fixes:
DNX fixes:
Temporarily revert fix for CS00012287482 - support for 1024 LAGs on DNX
CS00012297599 - [J2C+] sonic-mgmt failure in test_copp.py (test_no_policer[BGP])
CS00012293560 - ECN remark issue in SONiC
CS00012302371 - SONiC: V6 packets were mapped to wrong TC queue
CS00012288540 - Available ACL Entry and Counter is incorrect after removing ACL rules
Other changes (XGS fixes)
SID - L3 multicast packet drop due to wrong VFI derivation - SDK-350470
SID - SIGSEGV in linkscan callback delivery - SDK-287578
SID - Repeated VXLAN calls deletes vlan translation action profile SDK-313980
SER - error in IS_TDM_CALENDAR0/1 can cause traffic hit in TH
SID - L2_ENTRY Table Lookups May Miss
[CSP CS00012275452] sai_object_type_get_availability failed with SAI_STATUS_INVALID_PARAMETER
[CSP CS00012253527] sai_query_attribute_capability for obj type SAI_OBJECT_TYPE_SWITCH
Why I did it
When sonic is managed by k8s, the sonic container is managed by k8s daemonset, daemonset identifies its members by labels. Currently when restarting a sonic service by systemctl, if the service's container is already managed by k8s, systemd script stops the container by removing the feature label to make it disjoin from k8s daemonset, and then starts it by adding the label to make it join k8s daemonset again.
This behavior would cause problem during k8s container upgrade. Containers in daemonset are upgraded in a rolling fashion, that means the daemonset version is updated first, then rollout the new version to containers with precheck/postcheck one by one. However, if a sonic device joins a daemonset, k8s will directly deploy a pod with the current version of daemonset, it is expected when a device joins k8s cluster at first time.
But for a device which has already joined k8s cluster, the re-joining daemonset will cause the container upgraded to new version without precheck, so if a systemd service is restarted during daemonset upgrade, the container may be upgraded without precheck and break rolling update policy. To fix it, we need to remove the logic about dropping k8s label in systemd service stop script for kube mode.
Work item tracking
Microsoft ADO (number only): 24304563
How I did it
Don't drop label in systemd service stop script when feature's set_owner is kube. Only drop label when feature's set_owner is local.
How to verify it
The label feature_enabled should be always true if the feature's set owner is kube.
Modify snmpd.conf to start snmpd to listen on specific management and loopback ips instead of listening on any ip.
#### Why I did it
SNMP over IPv6 is not working for all scenarios for a single asic platforms.
The expectation is that SNMP query over IPv6 should work over Management or Loopback0 addresses.
**Specific scenario where this issue is seen**
In case of Lab T0 device, when SNMP request is sent from a directly connected T1 neighbor over Loopback IP, SNMP response was not received.
This was because the SRC IP address in SNMP response was not Loopback IP, it was the PortChannel IP connected to the neighboring device.
```
23:18:51.620897 In 22:26:27:e6:e0:07 ethertype IPv6 (0x86dd), length 105: fc00::72.41725 > **fc00:1::32**.161: C="msft" **GetRequest**(28) .1.3.6.1.2.1.1.1.0
23:18:51.621441 Out 28:99:3a:a0:97:30 ethertype IPv6 (0x86dd), length 241: **fc00::71**.161 > fc00::72.41725: C="msft" **GetResponse**(162) .1.3.6.1.2.1.1.1.0="SONiC Software Version: SONiC.xxx - HwSku: xx - Distribution: Debian 10.13 - Kernel: 4.19.0-12-2-amd64"
```
In case of IPv4, the SRC IP in SNMP response was correctly set to Loopback IP.
```
23:25:32.769712 In 22:26:27:e6:e0:07 ethertype IPv4 (0x0800), length 85: 10.0.0.57.56701 > **10.1.0.32**.161: C="msft" **GetRequest**(28) .1.3.6.1.2.1.1.1.0
23:25:32.975967 Out 28:99:3a:a0:97:30 ethertype IPv4 (0x0800), length 221: **10.1.0.32**.161 > 10.0.0.57.56701: C="msft" **GetResponse**(162) .1.3.6.1.2.1.1.1.0="SONiC Software Version: SONiC.xxx - HwSku: xx - Distribution: Debian 10.13 - Kernel: 4.19.0-12-2-amd64"
```
**Sequence of SNMP request and response**
1. SNMP request will be sent with SRC IP fc00::72 DST IP fc00:1::32
2. SNMP request is received at SONiC device is sent to snmpd which is listening on port 161 :::161/
3. snmpd process will parse the request create a response and sent to DST IP fc00::72.
snmpd process does not track the DST IP on which the SNMP request was received, which in this case is Loopback IP.
snmpd process will only keep track what is tht IP to which the response should be sent to.
4. snmpd process will send the response packet.
5. Kernel will do a route look up on destination IP and find the best path.
ip -6 route get fc00::72
fc00::72 from :: dev PortChannel101 proto kernel src fc00::71 metric 256 pref medium
5. Using the "src" ip from about, the response is sent out. This SRC ip is that of the PortChannel and not the device Loopback IP.
The same issue is seen when SNMP query is sent from a remote server over Management IP.
SONiC device eth0 --------- Remote server
SNMP request comes with SRC IP <Remote_server> DST IP <Mgmt IP>
If kernel finds best route to Remote_server_IP is via BGP neighbors, then it will send the response via front-panel interface with SRC IP as Loopback IP instead of Management IP.
Main issue is that in case of IPv6, snmpd ignores the IP address to which SNMP request was sent, in case of IPv6.
In case of IPv4, snmpd keeps track of DST IP of SNMP request, it will keep track if the SNMP request was sent to mgmt IP or Loopback IP.
Later, this IP is used in ipi_spec_dst as SRC IP which helps kernel to find the route based on DST IP using the right SRC IP.
https://github.com/net-snmp/net-snmp/blob/master/snmplib/transports/snmpUDPBaseDomain.c#L300
ipi.ipi_spec_dst.s_addr = srcip->s_addr
Reference: https://man7.org/linux/man-pages/man7/ip.7.html
```
If IP_PKTINFO is passed to sendmsg(2)
and ipi_spec_dst is not zero, then it is used as the local
source address for the routing table lookup and for
setting up IP source route options. When ipi_ifindex is
not zero, the primary local address of the interface
specified by the index overwrites ipi_spec_dst for the
routing table lookup.
```
**This issue is not seen on multi-asic platform, why?**
on multi-asic platform, there exists different network namespaces.
SNMP docker with snmpd process runs on host namespace.
Management interface belongs to host namespace.
Loopback0 is configured on asic namespaces.
Additional inforamtion on how the packet coming over Loopback IP reaches snmpd process running on host namespace: https://github.com/sonic-net/sonic-buildimage/pull/5420
Because of this separation of network namespaces, the route lookup of destination IP is confined to routing table of specific namespace where packet is received.
if packet is received over management interface, SNMP response also is sent out of management interface. Same goes with packet received over Loopback Ip.
##### Work item tracking
- Microsoft ADO **17537063**:
#### How I did it
Have snmpd listen on specific Management and Loopback IPs specifically instead of listening on any IP for single-asic platform.
Before Fix
```
admin@xx:~$ sudo netstat -tulnp | grep 161
udp 0 0 0.0.0.0:161 0.0.0.0:* 15631/snmpd
udp6 0 0 :::161 :::* 15631/snmpd
```
After fix
```
admin@device:~$ sudo netstat -tulnp | grep 161
udp 0 0 10.1.0.32:161 0.0.0.0:* 215899/snmpd
udp 0 0 10.3.1.1:161 0.0.0.0:* 215899/snmpd
udp6 0 0 fc00:1::32:161 :::* 215899/snmpd
udp6 0 0 fc00:2::32:161 :::* 215899/snmpd
```
**How this change helps with the issue?**
To see snmpd trace logs, modify snmpd to start using the below parameters, in supervisord.conf file
```
/usr/sbin/snmpd -f -LS0-7i -Lf /var/log/snmpd.log
```
When snmpd listens on any IP, snmpd binds to IPv4 and IPv6 sockets as below:
```
netsnmp_udpbase: binding socket: 7 to UDP: [0.0.0.0]:0->[0.0.0.0]:161
trace: netsnmp_udp6_transport_bind(): transports/snmpUDPIPv6Domain.c, 303:
netsnmp_udpbase: binding socket: 8 to UDP/IPv6: [::]:161
```
When IPv4 response is sent, it goes out of fd 7 and IPv6 response goes out of fd 8.
When IPv6 response is sent, it does not have the right SRC IP and it can lead to the issue described.
When snmpd listens on specific Loopback/Management IPs, snmpd binds to different sockets:
```
trace: netsnmp_udpipv4base_transport_bind(): transports/snmpUDPIPv4BaseDomain.c, 207:
netsnmp_udpbase: binding socket: 7 to UDP: [0.0.0.0]:0->[10.250.0.101]:161
trace: netsnmp_udpipv4base_transport_bind(): transports/snmpUDPIPv4BaseDomain.c, 207:
netsnmp_udpbase: binding socket: 8 to UDP: [0.0.0.0]:0->[10.1.0.32]:161
trace: netsnmp_register_agent_nsap(): snmp_agent.c, 1261:
netsnmp_register_agent_nsap: fd 8
netsnmp_udpbase: binding socket: 10 to UDP/IPv6: [fc00:1::32]:161
trace: netsnmp_register_agent_nsap(): snmp_agent.c, 1261:
netsnmp_register_agent_nsap: fd 10
netsnmp_ipv6: fmtaddr: t = (nil), data = 0x7fffed4c85d0, len = 28
trace: netsnmp_udp6_transport_bind(): transports/snmpUDPIPv6Domain.c, 303:
netsnmp_udpbase: binding socket: 9 to UDP/IPv6: [fc00:2::32]:161
```
When SNMP request comes in via Loopback IPv4, SNMP response is sent out of fd 8
```
trace: netsnmp_udpbase_send(): transports/snmpUDPBaseDomain.c, 511:
netsnmp_udp: send 170 bytes from 0x5581f2fbe30a to UDP: [10.0.0.33]:46089->[10.1.0.32]:161 on fd 8
```
When SNMP request comes in via Loopback IPv6, SNMP response is sent out of fd 10
```
netsnmp_ipv6: fmtaddr: t = (nil), data = 0x5581f2fc2ff0, len = 28
trace: netsnmp_udp6_send(): transports/snmpUDPIPv6Domain.c, 164:
netsnmp_udp6: send 170 bytes from 0x5581f2fbe30a to UDP/IPv6: [fc00::42]:43750 on fd 10
```
#### How to verify it
Verified on single asic and multi-asic devices.
Single asic SNMP query with Loopback
```
ARISTA01T1#bash snmpget -v2c -c xxx 10.1.0.32 1.3.6.1.2.1.1.1.0
SNMPv2-MIB::sysDescr.0 = STRING: SONiC Software Version: SONiC.xx - HwSku: Arista-7260xx - Distribution: Debian 10.13 - Kernel: 4.19.0-12-2-amd64
ARISTA01T1#bash snmpget -v2c -c xxx fc00:1::32 1.3.6.1.2.1.1.1.0
SNMPv2-MIB::sysDescr.0 = STRING: SONiC Software Version: SONiC.xx - HwSku: Arista-7260xxx - Distribution: Debian 10.13 - Kernel: 4.19.0-12-2-amd64
```
On multi-asic -- no change.
```
sudo netstat -tulnp | grep 161
udp 0 0 0.0.0.0:161 0.0.0.0:* 17978/snmpd
udp6 0 0 :::161 :::* 17978/snmpd
```
Query result using Loopback IP from a directly connected BGP neighbor
```
ARISTA01T2#bash snmpget -v2c -c xxx 10.1.0.32 1.3.6.1.2.1.1.1.0
SNMPv2-MIB::sysDescr.0 = STRING: SONiC Software Version: SONiC.xx - HwSku: xx - Distribution: Debian 9.13 - Kernel: 4.9.0-14-2-amd64
ARISTA01T2#bash snmpget -v2c -c xxx fc00:1::32 1.3.6.1.2.1.1.1.0
SNMPv2-MIB::sysDescr.0 = STRING: SONiC Software Version: SONiC.xx - HwSku: xx - Distribution: Debian 9.13 - Kernel: 4.9.0-14-2-amd64
```
<!--
If PR needs to be backported, then the PR must be tested against the base branch and the earliest backport release branch and provide tested image version on these two branches. For example, if the PR is requested for master, 202211 and 202012, then the requester needs to provide test results on master and 202012.
-->
Co-authored-by: SuvarnaMeenakshi <50386592+SuvarnaMeenakshi@users.noreply.github.com>
Why I did it
When do clean up container images, current code has two bugs need to be fixed. And some variables' name maybe cause confused, change the variables' name.
Work item tracking
Microsoft ADO (number only): 24502294
How I did it
We do clean up after tag latest successfully. But currently tag latest function only return 0 and 1, 0 means succeed and 1 means failed, when we get 1, we will retry, when we get 0, we will do clean up. Actually the code 0 includes another case we don't need to do clean up. The case is that when we are doing tag latest, the container image we want to tag maybe not running, so we can not tag latest and don't need to cleanup, we need to separate this case from 0, return -1 now.
When local mode(v1) -> kube mode(v2) happens, one problem is how to handle the local image, there are two cases. one case is that there was one kube v1 container dry-run(cause we don't relace the local if kube version = local version), we will remove the kube v1 image and tag the local version with ACR prefix and remove local v1 local tag. Another case is that there was no kube v1 container dry-run, we remove the local v1 image directly, cause the local v1 image should not be the last desire version.
About the docker_id variable, it may cause confused, it's actually docker image id, so rename the variable. About the two dicts and the list, rename them to be more readable.
How to verify it
Check tag latest and image clean up result.
Why I did it
During the upgrade process via k8s, the feature's systemd service will restart as well, all of the feature systemd service has restart number limit, and the limit number is too small, only three times. if fallback happens when upgrade, the start count will be 2, just once again, the systemd service will be down. So, need to bypass this. This restart function will be called when do local -> kube, kube -> kube, kube ->local, each time call this function, we indeed need to restart successfully, so do reset-failed every time we do restart.
When need to go back to local mode, we do systemd restart immediately without waiting the default restart interval time so that we can reduce the container down time.
Work item tracking
Microsoft ADO (number only):
24172368
How I did it
Before every restart for upgrade, do reset feature's restart number. The restart number will be reset to 0 to bypass the restart limit.
When need to go back to local mode, we do systemd restart immediately.
How to verify it
Feature's systemd service can be always restarted successfully during upgrade process via k8s.
src/sonic-platform-daemons
* bef58aa - (HEAD -> 202205, origin/202205) Added PCIe transaction check for all peripherals on the bus (#331) (10 hours ago) [Ashwin Srinivasan]
#### Why I did it
After k8s upgrade a container, k8s can only know the container is running, don't know the service's status inside container. So we need a probe inside container, k8s will call the probe to check whether the container is really ready.
##### Work item tracking
- Microsoft ADO **(number only)**: 22453004
#### How I did it
Add a health check probe inside config engine container, the probe will check whether the start service exit normally or not if the start service exists and call the python script to do container self-related specific checks if the script is there. The python script should be implemented by feature owner if it's needed.
more details: [design doc](https://github.com/sonic-net/SONiC/blob/master/doc/kubernetes/health-check.md)
#### How to verify it
Check path /usr/bin/readiness_probe.sh inside container.
#### Which release branch to backport (provide reason below if selected)
- [ ] 201811
- [ ] 201911
- [ ] 202006
- [ ] 202012
- [ ] 202106
- [ ] 202111
- [x] 202205
- [x] 202211
#### Tested branch (Please provide the tested image version)
- [x] 20220531.28
src/sonic-platform-common
* 56f227a - (HEAD -> 202205, origin/202205) More prevention of fatal exception caused by VDM dictionary missing fields when a transceiver has just been pulled (#376) (3 hours ago) [snider-nokia]
Why I did it
To reduce the container's dependency from host system
Work item tracking
Microsoft ADO (number only):
17713469
How I did it
Move the k8s container startup script to config engine container, other than mount it from host.
How to verify it
Check file path(/usr/share/sonic/scripts/container_startup.py) inside config engine container.
Signed-off-by: Yun Li <yunli1@microsoft.com>
Co-authored-by: Qi Luo <qiluo-msft@users.noreply.github.com>
How I did it
Free up Multiprocessing Manager resource at task stop request
[self.mpmgr.shutdown() in task_stop]
How to verify it
time systemctl stop system-health.service
* [Arista] Update hwsku.json for Arista-7050QX-32S-S4Q31
* Change to 3x10G(3)+1x1G(1) on Arista-7050QX-32S-S4Q31
Co-authored-by: byu343 <byu@arista.com>
* [chassis][lldp] Fix the lldp error log in host instance which doesn't contain front pannel ports
---------
Signed-off-by: mlok <marty.lok@nokia.com>
Co-authored-by: Marty Y. Lok <76118573+mlok-nokia@users.noreply.github.com>
* [buildsystem] Fix hiredis package version: 0.14.1-1 (#15461)
- Why I did it
To fix hiredis compilation
- How I did it
Changed package version: 0.14.0-3~bpo9+1 -> 0.14.1-1
- How to verify it
make configure PLATFORM=mellanox
make target/sonic-mellanox.bin
Signed-off-by: Nazarii Hnydyn <nazariig@nvidia.com>
* Update Makefile
---------
Signed-off-by: Nazarii Hnydyn <nazariig@nvidia.com>
Co-authored-by: Nazarii Hnydyn <nazariig@nvidia.com>
Why I did it
Fix incorrectly specified table name in the extra queues and extra pgs j2 files for 8101-32FH-O
How I did it
Update platform module to 202205.2.2.7
