Bug 1181097

Summary:	keepalived should start after network-online.target instead of network.target
Product:	[Fedora] Fedora	Reporter:	Artur <arus>
Component:	keepalived	Assignee:	Ryan O'Hara <rohara>
Status:	CLOSED CURRENTRELEASE	QA Contact:	Fedora Extras Quality Assurance <extras-qa>
Severity:	medium	Docs Contact:
Priority:	unspecified
Version:	21	CC:	bperkins, matthias, rohara
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2015-02-13 17:32:04 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 Artur 2015-01-12 11:19:11 UTC

Description of problem:

keepalived service is starting too early, before all interfaces have a chance to go up. Then in logs I see this:

Jan 12 11:54:30 r1 Keepalived_vrrp[853]: receive an invalid ip number count associated with VRID!
Jan 12 11:54:30 r1 Keepalived_vrrp[853]: bogus VRRP packet received on eno2 !!!
Jan 12 11:54:30 r1 Keepalived_vrrp[853]: VRRP_Instance(VIP_MGMT) ignoring received advertisment...

Version-Release number of selected component (if applicable):

keepalived-1.2.15-1.fc21.x86_64
Also affects previous versions of Fedora.

How reproducible:

Sometimes (too often), depends on how much time is needed to bring all interfaces up.

Steps to Reproduce:
1. reboot
2. check logs
3.

Actual results:

Jan 12 11:54:30 r1 Keepalived_vrrp[853]: receive an invalid ip number count associated with VRID!
Jan 12 11:54:30 r1 Keepalived_vrrp[853]: bogus VRRP packet received on eno2 !!!
Jan 12 11:54:30 r1 Keepalived_vrrp[853]: VRRP_Instance(VIP_MGMT) ignoring received advertisment...


Expected results:

All required IPs reported by "Netlink reflector" added.

Additional info:

I can fix it by modyfying systemd.service:
Wants=network-online.target
After=syslog.target network-online.target

Probably it could be also avoided by using preempt_delay setting, but I'm using nopreempt in my configuration (can't use both).

Comment 1 Ryan O'Hara 2015-01-12 14:53:52 UTC

Thanks for reporting this. I've never run into this problem myself, but I understand how it would happen. Just one question, why is it that you're seeing a "bogus VRRP packet" message prior to the interface(s) being up? I'm not making the connection.

Comment 2 Artur 2015-01-12 16:28:02 UTC

Those messages appear AFTER network is finally up. Sorry, I should put more steps to reproduce:

1. start master properly
2. reboot backup
3. check backup's logs
4. backup receives announcements from master and logs "bogus VRRP packet" every advert_int seconds.

I should also include earlier log entries:

Jan 12 10:20:08 r1 Keepalived_vrrp[853]: VRRP is trying to assign VIP to unknown vlan107 interface !!! go out and fix your conf !!!
Jan 12 10:20:08 r1 Keepalived_vrrp[853]: VRRP is trying to assign VIP to unknown vlan102 interface !!! go out and fix your conf !!!
(and so on)

I hope it is clear now.

Comment 3 Fedora Update System 2015-01-13 17:56:28 UTC

keepalived-1.2.15-2.fc20 has been submitted as an update for Fedora 20.
https://admin.fedoraproject.org/updates/keepalived-1.2.15-2.fc20

Comment 4 Fedora Update System 2015-01-13 17:56:34 UTC

keepalived-1.2.15-2.fc21 has been submitted as an update for Fedora 21.
https://admin.fedoraproject.org/updates/keepalived-1.2.15-2.fc21

Comment 5 Ryan O'Hara 2015-01-13 17:58:07 UTC

OK. I've build packages for F20, F21 and Rawhide. Feel free to test and update karma accordingly. Thanks!

Comment 6 Artur 2015-01-19 10:03:17 UTC

Karma updated. Thank You for the fast reaction.

Comment 7 Fedora Update System 2015-01-26 02:27:48 UTC

keepalived-1.2.15-2.fc21 has been pushed to the Fedora 21 stable repository.  If problems still persist, please make note of it in this bug report.

Comment 8 Fedora Update System 2015-01-26 02:30:05 UTC

keepalived-1.2.15-2.fc20 has been pushed to the Fedora 20 stable repository.  If problems still persist, please make note of it in this bug report.