[NETVIRT-286] Tunnel id update as a resault of ovs restart Created: 22/Nov/16  Updated: 07/Dec/16  Resolved: 07/Dec/16

Status: Resolved
Project: netvirt
Component/s: General
Affects Version/s: Boron
Fix Version/s: None

Type: Bug
Reporter: zan cohen Assignee: Olga Schukin
Resolution: Won't Do Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Operating System: All
Platform: All

Attachments: Text File Report.txt     Text File after_restart.txt     Text File before_restart.txt    
External issue ID: 7217

 Description   

Description:
************
1.Load Devstuck setup(no vm's configure).
2.Dump controler+compute ovs pipeline to file.(see before_restart.txt file).
3.ovs-restart to controller+compute ovs only
4.Dump controler+compute ovs pipeline to file.(see after_restart.txt file).

Defect
******
Diff for both before and after ovs restart - it can be seen that tunnel id change.

 Comments   
Comment by zan cohen [ 22/Nov/16 ]

Attachment before_restart.txt has been added with description: Before_ovs_restart

Comment by zan cohen [ 22/Nov/16 ]

Attachment after_restart.txt has been added with description: After_ovs_restart

Comment by zan cohen [ 22/Nov/16 ]

Attachment Report.txt has been added with description: pipe_line_comparison

Comment by Olga Schukin [ 07/Dec/16 ]

Restart of OVS where allocated NAPT switch causes re-installation of related rules for all computes. T36 hook you are concerned, have a match on tunnel-id which is just a label. The label is generated automatically from VPN+IP while hooks installation. Then rules are updated for all OVS.
Ref to look for the flow in karaf logs:
"NAT Service : Select a new NAPT switch ..."

      • Sum: works as designed. ***

Note:
I've paid attention to the timing. If there are vm's on the compute which is going to be restarted, then reset flows take a little more time and ID Manager succeeds to release a previous label. Therefore it will allocate same label(tunnel-id) for the VPN+IP after restart. In your case the new label was generated just because old label is not free still, see below:

IdManager | 358 - org.opendaylight.genius.idmanager-impl - 0.2.0.SNAPSHOT | Got pool IdLocalPool [poolName=vpnservices.2130706433, availableIds=AvailableIdHolder [low=70000, high=72999, cur=70004], releasedIds=ReleasedIdHolder [availableIdCount=1, timeDelaySec=30, delayedEntries=[

{Id: 70002 ReadyTimeSec: 1481115010}

]]]

Generated at Wed Feb 07 20:21:11 UTC 2024 using Jira 8.20.10#820010-sha1:ace47f9899e9ee25d7157d59aa17ab06aee30d3d.