Could you execute those commands and paste me the output? for node1:
/usr/local/zenloadbalancer/app/libexec/check_http -w 5 -c 5 -H 192.168.1.47 -u / -S -p 443 echo $? fo node2: /usr/local/zenloadbalancer/app/libexec/check_http -w 5 -c 5 -H 172.17.1.30 -u / -S -p 443 echo $? FarmGuardian analyzes the output error variable (echo $?) if your backends are responding with HTTP code 403 and check_http waits for a HTTP200 then the output error could be <> 0 and FuarmGuardian could mark it as down. I remember check_http could analyze the HTTP CODE *check_http --help* * This plugin will attempt to open an HTTP connection with the host.* * Successful connects return STATE_OK, refusals and timeouts return STATE_CRITICAL* * other errors return STATE_UNKNOWN. Successful connects, but incorrect reponse* * messages from the host result in STATE_WARNING return values * If the output error code $? is not 0 then it could be the reason of your issue. Regards! 2016-04-07 17:30 GMT+02:00 Jonathan Haddock < jonathan.hadd...@ict.ekservices.org>: > Hello, > > I've got an L4xNAT farm set up to load balance HTTPS traffic based on > priority. In testing I've disabled the primary (one with the smallest > priority value, according to the docs) web server and the site goes > offline. According to "back end status" the primary webserver is still > available - it certainly isn't given I've removed the network cable :). > > I've also tried removing the primary server and re-adding it, while the > server is disconnected, and the Zen still believes the the server is up: > [image: Inline images 1] > > > I've connected to the zen over SSH and run check_http, the server is > definitely down: > root@eks-dmz-nlb-03:~# /usr/local/zenloadbalancer/app/libexec/check_http > -w 5 -c 5 -H 172.17.1.30 -u / -S -p 443 > connect to address 172.17.1.30 and port 443: No route to host > HTTP CRITICAL - Unable to open TCP socket > > Setting the farm to use FarmGuardian (with the command set to "check_http > -w 5 -c 5 -H 172.17.1.30 -u / -S -p 443") causes both backends to be > considered down (the primary times out, the secondary gives an HTTP403, > which is fine from my perspective). > > If I change the cluster to "weight" mode it's then possible to access the > site via the secondary server (with the primary still unplugged). The Zen > still believes both backends are operational though. > > Configuration of the farm is shown in the screen shot is below, in case > it's useful. > > There are 2 Zen nodes in a cluster (both nodes may be master) running the > latest (3.10) community edition. > > Any thoughts appreciated, > > Jonathan > > [image: Inline images 2] > > Please note that I have a new email address > (*jonathan.hadd...@ict.ekservices.org > <jonathan.hadd...@ict.ekservices.org>*). Email is currently being > auto-forwarded, but will cease at some point, so please update your records > . > > *Jonathan Haddock * > *MSc MBCS*Network and Security Engineer > *EK Services* > > Tel No: 01227 862 036 > jonathan.hadd...@ict.ekservices.org > *EK Services - working in partnership with Canterbury City Council, Dover > District Council and Thanet District Council* > > > ------------------------------------------------------------------------------ > > _______________________________________________ > Zenloadbalancer-support mailing list > Zenloadbalancer-support@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support > > -- Load balancer distribution - Open Source Project http://www.zenloadbalancer.com Distribution list (subscribe): zenloadbalancer-support@lists.sourceforge.net
------------------------------------------------------------------------------
_______________________________________________ Zenloadbalancer-support mailing list Zenloadbalancer-support@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support
