Re: Problems with takeover even though master is online

Have you tried added -v to the pulse command line (in /etc/init.d/pulse)?  This 
will cause debugging messages to be written to the syslog, it may help to 
diagnose the problem.

It looks like everything was fine for two hours.  Hmm, this seems like some sort 
of timeout.  Is there anything at all unusual about your network layout?

I'd be curious to see the results of a tcpdump on port 539 (or whatever port 
you're heartbeating on) at two different times:
   - during the initial "working" period
   - after the backup takes over

Also, could you post your lvs.cf?

DI Peter Burgstaller wrote:
> Hi again,
> 
> I'm using piranha 0.7.0 and ipvsadm 1.20release6-2
> As stated before the loadbalancers are linux (Debian/woody) machines 
> with kernel 2.4.19
> 
> This is the syslog of both machines, the machines' system times are in 
> sync:
> After 19:00 it gets interesting...
> 
> syslog.5:Oct 10 17:04:50 lb01 pulse[13689]: STARTING PULSE AS MASTER
> syslog.5:Oct 10 17:14:05 lb01 pulse[13689]: partner dead: activating lvs
> syslog.5:Oct 10 17:14:10 lb01 pulse[13716]: gratuitous lvs arps finished
> syslog.5:Oct 10 19:12:07 lb01 pulse[13689]: partner active: resending arps
> 
> syslog.5:Oct 10 17:14:06 lb02 pulse[8567]: STARTING PULSE AS BACKUP
> syslog.5:Oct 10 19:12:05 lb02 pulse[8567]: partner dead: activating lvs
> syslog.5:Oct 10 19:12:10 lb02 pulse[8577]: gratuitous lvs arps finished
> syslog.5:Oct 10 19:12:12 lb02 pulse[8567]: partner active: deactivating lvs
> syslog.5:Oct 10 19:15:44 lb02 pulse[8567]: partner dead: activating lvs
> syslog.5:Oct 10 19:15:48 lb02 pulse[8567]: partner active: deactivating lvs
> syslog.5:Oct 10 19:15:49 lb02 pulse[8627]: gratuitous lvs arps finished
> syslog.5:Oct 10 19:41:50 lb02 pulse[8567]: partner dead: activating lvs
> syslog.5:Oct 10 19:41:54 lb02 pulse[8567]: partner active: deactivating lvs
> syslog.5:Oct 10 19:41:56 lb02 pulse[8675]: gratuitous lvs arps finished
> 
> As you can see .. the backup server thinks that the master is dead .. 
> even though
> it isn't.
> 
> Any advice?
> 
> Thanks, Peter
> 
> On Dienstag, Okt 15, 2002, at 22:50 Europe/Vienna, Mike McLean wrote:
> 
>> What version of piranha were you using?  The same version that Alex 
>> was (0.6.0-19)?
>>
>> Do the routers flip-flop, or does the backup just take over and remain 
>> active?
>>
>> Try adding -v to the pulse command line.
>>
>> DI Peter Burgstaller wrote:
>>
>>> Hi list,
>>> I'm using piranha on a debian (woody) system with 2 load balancers 
>>> and  2 web servers (nat). I have exactly the same problem as 
>>> described in
>>> https://listman.redhat.com/pipermail/piranha-list/2002-February/ 
>>> 000849.html
>>> Is there any solution to this problem?
>>> - Thanks, Peter
>>> _______________________________________________
>>> Piranha-list mailing list
>>> [email protected]
>>> https://listman.redhat.com/mailman/listinfo/piranha-list
>>
>>
>>
>>
>>
>> _______________________________________________
>> Piranha-list mailing list
>> [email protected]
>> https://listman.redhat.com/mailman/listinfo/piranha-list
> 
> 
> 
> 
> _______________________________________________
> Piranha-list mailing list
> [email protected]
> https://listman.redhat.com/mailman/listinfo/piranha-list