"xinetd: readjusting service" resulting in failover

marama

Hi folks.

Version 2.3 on ESXi on PcEngines APU.

This is what I have in my log:


2016-07-19T00:45:00+02:00 10.0.0.1 xinetd[30186]: readjusting service 19053-tcp
2016-07-19T00:45:00+02:00 10.0.0.1 xinetd[30186]: readjusting service 19053-udp
2016-07-19T00:45:00+02:00 10.0.0.1 xinetd[30186]: readjusting service 19054-tcp
2016-07-19T00:45:00+02:00 10.0.0.1 xinetd[30186]: readjusting service 19054-udp
2016-07-19T00:45:00+02:00 10.0.0.1 xinetd[30186]: readjusting service 19055-tcp
2016-07-19T00:45:00+02:00 10.0.0.1 xinetd[30186]: readjusting service 19055-udp
2016-07-19T00:45:00+02:00 10.0.0.1 xinetd[30186]: readjusting service 19056-tcp
2016-07-19T00:45:00+02:00 10.0.0.1 xinetd[30186]: readjusting service 19056-udp
2016-07-19T00:45:00+02:00 10.0.0.1 xinetd[30186]: Reconfigured: new=0 old=101 dropped=0 (services)
2016-07-19T00:45:01+02:00 10.0.0.1 php-cgi: rc.filter_configure_sync: MONITOR: GG_pref_UPC is down, omitting from routing group GW_UPC
2016-07-19T00:45:01+02:00 10.0.0.1 php-cgi: rc.filter_configure_sync: Message sent to monitoring@... OK
2016-07-19T00:45:01+02:00 10.0.0.1 php-cgi: rc.filter_configure_sync: MONITOR: GG_pref_Inode is down, omitting from routing group GW_UPC
2016-07-19T00:45:01+02:00 10.0.0.1 php-cgi: rc.filter_configure_sync: Message sent to monitoring@... OK
...

and it happens maybe once a day. As far as I can tell, it's a filter reloading triggering the GatewayGroup failover, and it's not something I wanted to have.
My gateway groups have a "Trigger Level" set to "Member down", with (I think) default failover thresholds:

Weight: 1
Data Payload: 0
Latency thresholds: 200/500
Packet Loss thresholds: 10/20
Probe Interval: 5000
Loss Interval: 2000
Time Period: 60000
Alert interval: 1000

So how can I avoid my setup to break WAN connection when such filter reloads occur?

Bye

cmb

It's not that the filter reload is causing a failover, the filter reload occurs because you're having a WAN failure. Likely that you are having loss, but why on two diff gateways at the same time isn't clear from that limited context. Check your gateway logs.