Bug 1200918

Summary: core dump ov openvswitch
Product: [Community] RDO Reporter: Miroslav Suchý <msuchy>
Component: openvswitchAssignee: Alan Pevec <apevec>
Status: CLOSED CURRENTRELEASE QA Contact: Ofer Blaut <oblaut>
Severity: low Docs Contact:
Priority: high    
Version: IcehouseCC: apevec, chrisw, madko, puiterwijk, yeylon
Target Milestone: ---Keywords: Reopened
Target Release: Kilo   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: openvswitch-2.3.1-2.el7 Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2015-04-11 20:10:13 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:

Description Miroslav Suchý 2015-03-11 16:02:55 UTC
Description of problem:
I have RDO Icehouse. I am preparing new OpenStack instance for Fedora infrastructure. I am very close to final state so I reprovisioned controller machine and installed it again using Ansible Playbook we created.
Unfortunately now, after I start OpenStack instance the ovswitch die with core dump. Sometimes even before the bridge is created. 

Version-Release number of selected component (if applicable):
openvswitch-2.1.2-2.el7.centos.1.x86_64

How reproducible:
Never happened before, but now it die every time I try.

Steps to Reproduce:
1. install controller using this playbook
https://infrastructure.fedoraproject.org/cgit/ansible.git/tree/playbooks/hosts/fed-cloud09.cloud.fedoraproject.org.yml

Actual results:
openvswitch will die when OS services are started

Expected results:
no errors

Additional info:
This is the traceback:
#0  nl_attr_get_size (nla=nla@entry=0x0) at lib/netlink.c:516
#1  0x00007fbbe335eab3 in format_generic_odp_key (a=a@entry=0x0, ds=ds@entry=0x7fbbd7ffcd00) at lib/odp-util.c:747
#2  0x00007fbbe335f5b0 in format_odp_key_attr (a=a@entry=0x7fbbb400094c, ma=ma@entry=0x0, portno_names=portno_names@entry=0x0, 
    ds=ds@entry=0x7fbbd7ffcd00, verbose=verbose@entry=true) at lib/odp-util.c:1369
#3  0x00007fbbe335f27e in odp_flow_format (key=key@entry=0x7fbbb400094c, key_len=key_len@entry=148, mask=mask@entry=0x0, 
    mask_len=mask_len@entry=0, portno_names=portno_names@entry=0x0, ds=ds@entry=0x7fbbd7ffcd00, verbose=verbose@entry=true) at lib/odp-util.c:1440
#4  0x00007fbbe3341335 in log_flow_message (error=error@entry=2, operation=operation@entry=0x7fbbe33de22f "flow_get", 
    key=key@entry=0x7fbbb400094c, key_len=key_len@entry=148, mask=mask@entry=0x0, mask_len=mask_len@entry=0, stats=stats@entry=0x7fbb88000d48, 
    actions=0x0, actions_len=0, dpif=0x7fbbe50f7580) at lib/dpif.c:1500
#5  0x00007fbbe33417cd in dpif_flow_get (dpif=0x7fbbe50f7580, key=0x7fbbb400094c, key_len=148, actionsp=actionsp@entry=0x7fbbd7ffce48, 
    stats=stats@entry=0x7fbb88000d48) at lib/dpif.c:861
#6  0x00007fbbe3322327 in revalidate_ukey (ukey=0x7fbbb4000900, udump=0x7fbb88000c10, udpif=0x7fbbe510b300) at ofproto/ofproto-dpif-upcall.c:1320
#7  revalidate_udumps (revalidator=revalidator@entry=0x7fbbe5113ec8, udumps=udumps@entry=0x7fbbd7ffec00) at ofproto/ofproto-dpif-upcall.c:1554
#8  0x00007fbbe33228ab in udpif_revalidator (arg=0x7fbbe5113ec8) at ofproto/ofproto-dpif-upcall.c:727
#9  0x00007fbbe285cdf5 in start_thread () from /lib64/libpthread.so.0
#10 0x00007fbbe20801ad in clone () from /lib64/libc.so.6

Comment 1 Miroslav Suchý 2015-03-11 16:05:56 UTC
CoreDump [37MB]
https://msuchy.fedorapeople.org/core.1500

Comment 2 Patrick Uiterwijk 2015-03-11 17:15:10 UTC
This has been fixed after upgrading openvswitch to 2.3

Comment 3 Alan Pevec 2015-03-26 12:22:13 UTC
(In reply to Patrick Uiterwijk from comment #2)
> This has been fixed after upgrading openvswitch to 2.3

Was that https://cbs.centos.org/koji/buildinfo?buildID=568 ?

We still need to ship this in active RDO repos, it was pending verification in bug 1193429

BTW openvswitch-2.1.2-2.el7_0.1 is incompatible with RHEL 7.1 kernel but should work with 7.0 kernel.

Comment 4 Alan Pevec 2015-03-30 16:50:30 UTC
RDO Juno update https://review.gerrithub.io/228852