sonic-net / sonic-buildimage

Scripts which perform an installable binary image build for SONiC
Other
735 stars 1.41k forks source link

OA apply view failed if bridge_port type tunnel is configured #11167

Open msosyak opened 2 years ago

msosyak commented 2 years ago

Description

On a system, with up and running VxLAN L2EVPN tunnel after warm-reboot OA can't apply view with the following logs


Jun  9 09:00:21.649822 sonic ERR syncd[25]: :- getSaiAttrFromDefaultValue: serialization type SAI_ATTR_VALUE_TYPE_IP_ADDRESS is not supported yet, FIXME
Jun  9 09:00:21.652726 sonic ERR syncd[25]: message repeated 2 times: [ :- getSaiAttrFromDefaultValue: serialization type SAI_ATTR_VALUE_TYPE_IP_ADDRESS is not supported yet, FIXME]
Jun  9 09:00:21.652726 sonic WARNING syncd[25]: :- performObjectSetTransition: current attr is CREATE_ONLY and object is MATCHED: oid:0x3a000000000441 transferring SAI_BRIDGE_PORT_ATTR_TUNNEL_ID:oid:0x2a00000000043f to temp object
Jun  9 09:00:21.652726 sonic ERR syncd[25]: :- performObjectSetTransition: current attribute is mandatory on create, crate and set, and object MATCHED, FIXME oid:0x3a000000000441 SAI_BRIDGE_PORT_ATTR_BRIDGE_ID:oid:0x39000000000001
Jun  9 09:00:21.652726 sonic ERR syncd[25]: :- processObjectForViewTransition: performObjectSetTransition on MATCHED object (oid:0x3a000000000441) FAILED! bug?
Jun  9 09:00:21.652816 sonic NOTICE syncd[25]: :- applyViewTransition: comparison logic took 0.009119 sec
Jun  9 09:00:21.657236 sonic ERR syncd[25]: :- applyView: Exception: :- processObjectForViewTransition: performObjectSetTransition on MATCHED object (oid:0x3a000000000441) FAILED! bug?
Jun  9 09:00:21.658135 sonic NOTICE syncd[25]: :- applyView: apply took 0.078622 sec
Jun  9 09:00:21.659148 sonic ERR swss#orchagent: :- syncd_apply_view: Failed to notify syncd APPLY_VIEW -1

Steps to reproduce the issue:

  1. Configure VxLAN L2EVPN tunnel
  2. run warm-reboot

Describe the results you received:

OA failed at the first iteration.

Describe the results you expected:

OA start successfully at the first time

Output of show version:

SONiC Software Version: SONiC.master.88058-dirty-20220407.201045
Distribution: Debian 11.3
Kernel: 5.10.0-8-2-amd64
Build commit: 16717d2dc
Build date: Thu Apr  7 20:17:47 UTC 2022
Built by: AzDevOps@sonic-build-workers-001CZE

Platform: x86_64-accton_as9516_32d-r0
HwSKU: newport
ASIC: barefoot
ASIC Count: 1
Serial Number: 9516D2119028
Model Number: NP5ZZ8632007A
Hardware Revision: N/A
Uptime: 17:10:31 up 15 min,  1 user,  load average: 0.33, 0.46, 0.49
Date: Thu 16 Jun 2022 17:10:31

Output of show techsupport:

(paste your output here or download and attach the file here )

Additional information you deem important (e.g. issue happens only occasionally):

cold-reboot log files: sairedis.rec.txt swss.rec.txt syslog.log warm-reboot log files: wr-sairedis.rec.txt wr-swss.rec.txt wr-syslog.log

msosyak commented 2 years ago

@kcudnik Could you, please, have a look.

msosyak commented 2 years ago

@kcudnik I see that you added this transition logic and the SAI_BRIDGE_PORT_ATTR_BRIDGE_ID is not supported here yet https://github.com/Azure/sonic-sairedis/blob/4b2638cab69f245cc36122f37d6c88bfc7999b86/syncd/ComparisonLogic.cpp#L1602. But I do not really understand why. Could you help me with that?

kcudnik commented 2 years ago

do you have sairedis.rec from this warm boot ? i will need full recording from cold boot and warm boot

kcudnik commented 2 years ago

this is some corner case, i will need that recording to reproduce that locally then i could start looking why this is failing

msosyak commented 2 years ago

I have attached sairedis and other log files as well. Thank you in advance.

saiarcot895 commented 2 years ago

Would it be possible to attach the VxLAN L2EVPN part of your config (or the steps to configure it) to make it easier to locally test it?

saiarcot895 commented 2 years ago

Also, what version of SONiC did the upgrade happen from? I can see that the version after upgrade is SONiC-OS-master.88058-dirty-20220407.201045.

msosyak commented 2 years ago

My setup consists of two directly connected devices(one direct link), but the same could be configured on any two devices with an IP connection. Here are the conf files for both devices. LEAF1-l2evpn.json.txt LEAF2-l2evpn.json.txt Also after loading this configuration the following vtysh command should be done:

msosyak commented 2 years ago

Also, what version of SONiC did the upgrade happen from? I can see that the version after upgrade is SONiC-OS-master.88058-dirty-20220407.201045.

There is no upgrade during warmreboot. The SONiC version remains unchanged

msosyak commented 2 years ago

@saiarcot895 Were you able to reproduce the issue? Please, let me know if you need more information or maybe some debug session.

msosyak commented 2 years ago

@saiarcot895 @kcudnik Is there any update on this?

msosyak commented 2 years ago

Are there any updates?

kcudnik commented 2 years ago

@saiarcot895 any updates ?

saiarcot895 commented 2 years ago

No updates yet, still working on this.

msosyak commented 2 years ago

@saiarcot895 Have you had a chance to look at this?

saiarcot895 commented 2 years ago

Hi @msosyak,

I have some code changes that might fix this, but when running with a recent image, it seems that there's been some changes done to VXLAN tunnel creation codepath. Could you recheck to see if there's still a warm reboot issue with this particular config?