[Keepalived-devel] Two MASTERS at the same time after FAULT state

[ROHAUT S. (E. G. <ext...@sn...>]

Hi,

 

We are using Keepalived 1.1.20 on Debian 5 servers. We are encountering a problem with two servers in MASTER state.

 

We have two keepalived servers, one initially master, one initially backup. The first server encountered a link down on one network interface and all VRRP instances commuted to FAULT state (vrrp_sync_group). The second server commuted to MASTER state as expected.

 

After a few seconds, the network interface on first server goes up again, and it commuted to MASTER state. Problem : VRRP instance has nopreempt so the first server should switch to Backup state. And the second server remains MASTER ! First server losts his link a second time, commuted again in FAULT state, then in MASTER state, but second server remains MASTER.

 

So we now have two MASTERS, with duplicate IP on both servers.

 

>From server1 I can ping Server2 without any problem.

 

Thanks for your help.

 

Condensed logs en server1 (Initial MASTER) :

 

Jul  8 09:38:33 server1 kernel: [70184.545105] tg3: eth2: Link is down.

Jul  8 09:38:33 server1 Keepalived_vrrp: VRRP_Group(VG1) Syncing instances to FAULT state

Jul  8 09:38:48 server1 kernel: [70199.148036] tg3: eth2: Link is up at 1000 Mbps, full duplex.

Jul  8 09:38:48 server1 Keepalived_vrrp: VRRP_Group(VG1) Syncing instances to BACKUP state

Jul  8 09:38:50 server1 Keepalived_vrrp: VRRP_Group(VG1) Syncing instances to MASTER state

Jul  8 09:38:52 server1 kernel: [70203.475639] tg3: eth2: Link is down.

Jul  8 09:38:52 server1 Keepalived_vrrp: VRRP_Group(VG1) Syncing instances to FAULT state

Jul  8 09:39:14 server1 kernel: [70225.095566] tg3: eth2: Link is up at 100 Mbps, full duplex.

Jul  8 09:39:14 server1 Keepalived_vrrp: VRRP_Group(VG1) Syncing instances to BACKUP state

Jul  8 09:39:16 server1 Keepalived_vrrp: VRRP_Group(VG1) Syncing instances to MASTER state

 

 

Logs en Server2 (initial BACKUP) :

Jul  8 09:38:36 server2 Keepalived_vrrp: VRRP_Instance(VI_1) Transition to MASTER STATE

Jul  8 09:38:36 server2 Keepalived_vrrp: VRRP_Group(VG1) Syncing instances to MASTER state

Jul  8 09:38:36 server2 Keepalived_vrrp: VRRP_Instance(VI_2) Transition to MASTER STATE

Jul  8 09:38:37 server2 Keepalived_vrrp: VRRP_Instance(VI_2) Entering MASTER STATE

Jul  8 09:38:37 server2 Keepalived_vrrp: VRRP_Instance(VI_1) Entering MASTER STATE

 

vrrp_sync_group VG1 {

        group {

                VI_1

                VI_2

        }

}

 

vrrp_instance VI_1 {

        state BACKUP

        interface eth2

        priority 150

        advert_int 1

        nopreempt

}

 

vrrp_instance VI_2 {

        state BACKUP

        interface eth1

        priority 150

        advert_int 1

        nopreempt

}

 

Cordialement,

 

Sébastien ROHAUT

-------
Ce message et toutes les pièces jointes sont établis à l'intention exclusive de ses destinataires et sont confidentiels. L'intégrité de ce message n'étant pas assurée sur Internet, la SNCF ne peut être tenue responsable des altérations qui pourraient se produire sur son contenu. Toute publication, utilisation, reproduction, ou diffusion, même partielle, non autorisée préalablement par la SNCF, est strictement interdite. Si vous n'êtes pas le destinataire de ce message, merci d'en avertir immédiatement l'expéditeur et de le détruire.
-------
This message and any attachments are intended solely for the addressees and are confidential. SNCF may not be held responsible for their contents whose accuracy and completeness cannot be guaranteed over the Internet. Unauthorized use, disclosure, distribution, copying, or any part thereof is strictly prohibited. If you are not the intended recipient of this message, please notify the sender immediately and delete it.