[
https://issues.apache.org/jira/browse/CLOUDSTACK-9015?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14986865#comment-14986865
]
Wilder Rodrigues commented on CLOUDSTACK-9015:
----------------------------------------------
Hi [~giraffeforestg],
I know how to make it work with workarounds as well: just restarting keepalived
in one of the routers and it will be fine.
What I found out is that keepalived is being started before the interfaces are
configured, that’s why there is no advertisement and the thing simply doesn’t
work. if we start it after the interfaces are fine, it works.
In addition, it works fine with a stop/start, instead of a reboot. there is a
step that the router does differently when rebooting. And I'm 100% sure it
works and won't even look at the stop/start case because the
component/test_vpc_redundant.py already does it. That test has been executed at
leat 50 times with all the PRs that have been merged in the last month.
Probably the default allow not being used means that keepalived starts, but
since there is no outgoing traffic, its state gets right. When there is
outgoing traffic, it breaks because the interfaces are not configured, on both
sides!
I will try a fix when get to the office. On the "set_redundant_on" code, I will
wait for the interface to become available. Will give it a try and will let you
know.
Thanks for your reports!
Cheers,
Wilder
> Redundant VPC Virtual Router's state is BACKUP & BACKUP
> -------------------------------------------------------
>
> Key: CLOUDSTACK-9015
> URL: https://issues.apache.org/jira/browse/CLOUDSTACK-9015
> Project: CloudStack
> Issue Type: Bug
> Security Level: Public(Anyone can view this level - this is the
> default.)
> Affects Versions: 4.6.0
> Environment: CloudStack master(2015/10/31) 4.6.0-snapshot
> Hypervisor CentOS6/KVM
> SystemVM
> build #654 (2015/10/22 19:27:55)
> http://jenkins.buildacloud.org/job/build-systemvm64-master/lastSuccessfulBuild/artifact/tools/appliance/dist/systemvm64template-master-4.6.0-kvm.qcow2.bz2
> Reporter: satoru nakaya
> Assignee: Wilder Rodrigues
> Priority: Critical
>
> Steps of reproduce.
> 1)Create VPC (Redundant VPC offering)
> 2)Create tier
> 3)Create VM Instance on this tier
> 4)Check Redundant state (good)
> r-14-VM Redundant state:MASTER
> r-15-VM Redundant state:BACKUP
> 5) Reboot Router r-14-VM
> 6)Check Redundant state (good)
> r-14-VM Redundant state:BACKUP
> r-15-VM Redundant state:MASTER
> 7) Reboot Router r-15-VM
> 8)Check Redundant state (bad)
> r-14-VM Redundant state:BACKUP
> r-15-VM Redundant state:BACKUP
> 9)Check Log(r-14-VM's /var/log/messages)
> Nov 1 00:46:29 r-14-VM Keepalived_vrrp[3711]: VRRP_Instance(inside_network)
> sending 0 priority
> Nov 1 00:47:34 r-14-VM Keepalived_vrrp[2179]: Netlink: filter function error
> Nov 1 00:47:34 r-14-VM Keepalived_healthcheckers[2178]: Netlink: filter
> function error
> Nov 1 00:47:53 r-14-VM Keepalived_healthcheckers[2178]: Netlink: filter
> function error
> Nov 1 00:47:53 r-14-VM Keepalived_vrrp[2179]: Netlink: filter function error
> Nov 1 00:47:54 r-14-VM Keepalived_healthcheckers[2178]: Netlink: filter
> function error
> Nov 1 00:47:54 r-14-VM Keepalived_vrrp[2179]: Netlink: filter function error
> 10)Check Log(r-15-VM's /var/log/messages)
> Nov 1 00:49:19 r-15-VM Keepalived_vrrp[3682]: VRRP_Instance(inside_network)
> sending 0 priority
> Nov 1 00:50:25 r-15-VM Keepalived_vrrp[2179]: Netlink: filter function error
> Nov 1 00:50:25 r-15-VM Keepalived_healthcheckers[2178]: Netlink: filter
> function error
> Nov 1 00:50:45 r-15-VM Keepalived_healthcheckers[2178]: Netlink: filter
> function error
> Nov 1 00:50:45 r-15-VM Keepalived_vrrp[2179]: Netlink: filter function error
> Nov 1 00:50:45 r-15-VM Keepalived_healthcheckers[2178]: Netlink: filter
> function error
> Nov 1 00:50:45 r-15-VM Keepalived_vrrp[2179]: Netlink: filter function error
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
