TCP packets getting lost causing same-subnet-tcp-connections to hang
Bug #1393730 reported by
Vedamurthy Joshi
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Juniper Openstack |
Fix Committed
|
High
|
Anand H. Krishnan |
Bug Description
Build 2.0 2464 Single node setup
nodeb8 is a single node sanity setup
After provisioning, it was seen that machines on the same subnet as the compute node fail to establish ssh sessions to this compute node.
Machines outside of the subnet are able to ssh fine
Anand would have more details once his debugging is complete
Changed in juniperopenstack: | |
status: | New → Fix Committed |
tags: | added: sanity |
To post a comment you must log in.
Reviewed: https:/ /review. opencontrail. org/4705 github. org/Juniper/ contrail- vrouter/ commit/ c841170383f491e 670b10ca998d4c3 5dcba9283c
Committed: http://
Submitter: Zuul
Branch: master
commit c841170383f491e 670b10ca998d4c3 5dcba9283c
Author: Anand H Krishnan <email address hidden>
Date: Tue Nov 18 02:17:13 2014 -0800
Change the location from where we try to pull the outer headers
The function that pulls outer headers assumes that the packet data
points at layer 2 header. While this is true in case of ubuntu, the
commit that made this assumption did not reposition the call for
Centos, resulting in pull failures.
Change-Id: Iaae280817820fe b87a525db8c028b cba4e905bb1
Closes-BUG: #1393730