[BGPCEP-876] Controller stuck learning 300K routes from 2 peers Created: 19/Jul/19  Updated: 24/Jul/19  Resolved: 24/Jul/19

Status: Resolved
Project: bgpcep
Component/s: BGP
Affects Version/s: Sodium, Fluorine SR3, Neon SR2
Fix Version/s: Neon SR1, Sodium, Fluorine SR3

Type: Bug Priority: High
Reporter: Luis Gomez Assignee: Robert Varga
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified


 Description   

Regression was detected here:

https://jenkins.opendaylight.org/releng/view/bgpcep/job/bgpcep-csit-1node-bgp-ingest-all-neon/

The test executes this:

python play.py --amount=300000 --multiplicity=2 --myip=127.0.0.1 --myport=17900 --peerip=10.30.170.74 --peerport=1790 --logfile=bgp_peer.log --info &> play.py.out

For the test to pass we had lower down from 300K down to 100K:

https://git.opendaylight.org/gerrit/#/c/83159/



 Comments   
Comment by Robert Varga [ 23/Jul/19 ]

I think this is related to two things:

  • the change in play.py to log reading
  • something in infra

i.e. https://git.opendaylight.org/gerrit/#/q/status:merged+branch:stable/fluorine is showing absolutely nothing, yet https://jenkins.opendaylight.org/releng/view/bgpcep/job/bgpcep-csit-1node-bgp-ingest-all-fluorine/ shows start of intermittent failures on July 4 – now being in a place where every other run succeeded. -neon and -sodium both are showing a similar pattern, but are failing pretty consistently since July 12, when the play.py change was merged.

Comment by Luis Gomez [ 23/Jul/19 ]

The first actually impacted, the revert patch seems to work: https://git.opendaylight.org/gerrit/c/integration/test/+/83202, we can close the ticket for now.

Comment by Luis Gomez [ 23/Jul/19 ]

This is the revert to restore 300K routes: https://git.opendaylight.org/gerrit/c/integration/test/+/83217

Generated at Wed Feb 07 19:14:22 UTC 2024 using Jira 8.20.10#820010-sha1:ace47f9899e9ee25d7157d59aa17ab06aee30d3d.