Bug 1739797

Summary: [Regression] Hi CPU usage caused by RA flood
Product: [Fedora] Fedora Reporter: Vladimir Benes <vbenes>
Component: dnsmasqAssignee: Petr Menšík <pemensik>
Status: CLOSED ERRATA QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: urgent Docs Contact:
Priority: urgent    
Version: 31CC: bugzilla, code, dougsland, extras-qa, fge, itamar, jima, laine, p, pemensik, thozza, veillard
Target Milestone: ---Keywords: Regression, TestCaseProvided
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: dnsmasq-2.80-10.fc31 dnsmasq-2.80-10.fc30 Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: 1739081 Environment:
Last Closed: 2019-09-14 00:06:49 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On: 1739081    
Bug Blocks:    

Description Vladimir Benes 2019-08-10 14:59:19 UTC
I see a huge flood in journalctl from NM, which is probably caused by tons of RA announcements.

Aug 10 10:55:28 testsethostname dnsmasq-dhcp[18631]: RTR-ADVERT(simbr) 2620:52:0:1086::
Aug 10 10:55:28 testsethostname NetworkManager[21459]: <debug> [1565448928.5546] ndisc-lndp[0x55cae74930e0,"eth0"]: processing libndp events
Aug 10 10:55:28 testsethostname dnsmasq-dhcp[18631]: RTR-ADVERT(simbr) 2620:52:0:1086::
Aug 10 10:55:28 testsethostname NetworkManager[21459]: <debug> [1565448928.5548] ndisc-lndp[0x55cae74930e0,"eth0"]: processing libndp events
Aug 10 10:55:28 testsethostname dnsmasq-dhcp[18631]: RTR-ADVERT(simbr) 2620:52:0:1086::
Aug 10 10:55:28 testsethostname NetworkManager[21459]: <debug> [1565448928.5551] ndisc-lndp[0x55cae74930e0,"eth0"]: processing libndp events
Aug 10 10:55:28 testsethostname dnsmasq-dhcp[18631]: RTR-ADVERT(simbr) 2620:52:0:1086::
Aug 10 10:55:28 testsethostname NetworkManager[21459]: <debug> [1565448928.5553] ndisc-lndp[0x55cae74930e0,"eth0"]: processing libndp events


simbr is configured in virtual setup environment used for NetworkManager's tests.
    # Create the 'simbr' - providing both 10.x ipv4 and 2620:52:0 ipv6 dhcp
    ip netns exec vethsetup ip link add name simbr type bridge forward_delay 0 stp_state 1
    ip netns exec vethsetup ip link set simbr up
    ip netns exec vethsetup ip addr add 10.16.1.1/24 dev simbr
    ip netns exec vethsetup ip -6 addr add 2620:52:0:1086::1/64 dev simbr

    # Add eth10 peer into the simbr
    ip netns exec vethsetup ip link set eth10p master simbr

    # Run joint DHCP4/DHCP6 server with router advertisement enabled in veth namespace
    ip netns exec vethsetup dnsmasq --pid-file=/tmp/dhcp_simbr.pid --dhcp-leasefile=/tmp/dhcp_simbr.lease --dhcp-range=10.16.1.10,10.16.1.254,240 --dhcp-range=2620:52:0:1086::10,2620:52:0:1086::1ff,slaac,64,240 --enable-ra --interface=simbr --bind-interfaces

Comment 1 Ben Cotton 2019-08-13 17:04:46 UTC
This bug appears to have been reported against 'rawhide' during the Fedora 31 development cycle.
Changing version to '31'.

Comment 2 Ben Cotton 2019-08-13 17:10:22 UTC
This bug appears to have been reported against 'rawhide' during the Fedora 31 development cycle.
Changing version to 31.

Comment 3 Petr Menšík 2019-08-21 17:54:40 UTC
Confirming bug, it is in upstream. Created restartable reproducer [1]. Reliably reports many packets on v2.80, but v2.79 is ok. Vanilla build from upstream is affected too.

1. https://github.com/InfrastructureServices/dnsmasq-tests/blob/master/bz1739797.sh

Comment 4 Petr Menšík 2019-08-21 18:54:11 UTC
Found breaking commit to be [1]. It seems to be intention to send it fast, but maybe damage created was not well considered. Impact is significant.

1. http://thekelleys.org.uk/gitweb/?p=dnsmasq.git;a=commit;h=0a496f059c1e9d75c33cce4c1211d58422ba4f62

Comment 5 Petr Menšík 2019-08-21 20:30:50 UTC
Sent information upstream, for now without any patch.

http://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2019q3/013255.html

Comment 6 Fedora Update System 2019-08-28 17:56:03 UTC
FEDORA-2019-7f297cb49a has been submitted as an update to Fedora 30. https://bodhi.fedoraproject.org/updates/FEDORA-2019-7f297cb49a

Comment 7 Fedora Update System 2019-08-28 17:56:46 UTC
FEDORA-2019-883c54f9e9 has been submitted as an update to Fedora 31. https://bodhi.fedoraproject.org/updates/FEDORA-2019-883c54f9e9

Comment 8 Fedora Update System 2019-08-30 00:04:39 UTC
dnsmasq-2.80-10.fc30 has been pushed to the Fedora 30 testing repository. If problems still persist, please make note of it in this bug report.
See https://fedoraproject.org/wiki/QA:Updates_Testing for
instructions on how to install test updates.
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2019-7f297cb49a

Comment 9 Fedora Update System 2019-08-30 12:17:07 UTC
dnsmasq-2.80-10.fc31 has been pushed to the Fedora 31 testing repository. If problems still persist, please make note of it in this bug report.
See https://fedoraproject.org/wiki/QA:Updates_Testing for
instructions on how to install test updates.
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2019-883c54f9e9

Comment 10 Fedora Update System 2019-09-14 00:06:49 UTC
dnsmasq-2.80-10.fc31 has been pushed to the Fedora 31 stable repository. If problems still persist, please make note of it in this bug report.

Comment 11 Fedora Update System 2019-09-14 01:12:12 UTC
dnsmasq-2.80-10.fc30 has been pushed to the Fedora 30 stable repository. If problems still persist, please make note of it in this bug report.

Comment 12 Fedora Update System 2019-09-14 16:31:01 UTC
dnsmasq-2.80-10.fc31 has been pushed to the Fedora 31 stable repository. If problems still persist, please make note of it in this bug report.