Re: Real Server taken back into service after restart (here: Apache)

Hi, Bernd!

Yes, this is intended. Keepalived had removed your server from ipvs when it fell down, and then re-created it when it became available for checks with the configured weight=1.

Consider using the "inhibit_on_failure" config option to make the keepalived do things you tried to do with ipvsadm.


19.06.2015 12:49, Bernd пишет:
> Hi,
> 
> I'm running keepalived-1.2.16-5.x86_64 (self-built via rpmbuild from 
> your sources) on 3.10.0-229.4.2.el7.x86_64 (CentOS 7).
> 
> Recently, I wanted to take out one of three backend servers (serving 
> http/https via Apache). I issued
> 
> ipvsadm -e -t 123.123.123.123:80 -r 10.11.12.13:80 -m -w 0
> ipvsadm -e -t 123.123.123.123:443 -r 10.11.12.13:443 -m -w 0
> 
> Server was sessionless after some time 
> (net.ipv4.vs.expire_quiescent_template = 0).
> 
> Maintenance on the server took place, which included three times restart 
> of Apache.
> 
> However, keepalived was triggered by this to take the machine *back into 
> service*, read: It set weight itself back to '1'.
> 
> Is that behaviour intended?
> 
> Best,
> 
> Bernd
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> Keepalived-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/keepalived-devel
> 

-- 
Best regards,
Alexey 

------------------------------------------------------------------------------
_______________________________________________
Keepalived-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/keepalived-devel