Hi, Your request is legitimate, but I don't see how to do this with HAProxy itself. You need to write a third party script to do this.
What I would do is to configure HAProxy logs, configure my syslog to isolate legitimate traffic from server status changes. That way, with a third party script it's easy to know if a server is flapping or not. Then you could disable the flapping server using the haproxy stats socket. cheers On Thu, Feb 2, 2012 at 9:24 PM, John Clegg <[email protected]> wrote: > That's the problem . We did have a hardware issue on the NIC and the load > balancer kept putting the server live / dead all the time. Unfortunately the > NIC wasn't monitored so we had no idea the NIC was flapping. (This has is > now monitored) > > My issue that the NIC flapping caused the load balancer to flap and that > caused a lot of customer issues which impacted revenue for the business :-( > It would have been better for that server to have marked dead and the > servers under the load balancer would have taken over. > > My reasoning is that if there is a weird hardware / server fault or its > flapping for too long / too many times, I want the load balancer to mark the > server dead. I can then pick that up with my monitoring. > > Or is there a better strategy? > > John > > > On Fri, Feb 3, 2012 at 1:43 AM, Baptiste <[email protected]> wrote: >> >> Hey, >> >> According to me, it's easier to change the NIC, since your >> intermittent issues may happen under load only. >> >> With HAProxy, I can't see anything you can currently do to avoid this >> behavior. >> I mean that NIC intermittent issues may happen under load, so haproxy >> would consider server down. >> load decreasing, NIC will work properly again and haproxy would >> consider server operational. >> and so on for a lonnnnnnnnnng time. >> >> Don't try to hide a hardware issue using a load-balancer... >> >> cheers >> >> >> On Thu, Feb 2, 2012 at 4:28 AM, John Clegg <[email protected]> wrote: >> > Hi >> > >> > I'm trying to understand how ensure backserver which is failing and >> > classified as dead stays dead. >> > >> > I've just had an instance on another server which is using another >> > load-balancer where the NIC has intermittently failing and it caused the >> > load-balancer to flap constantly. >> > >> > I would like to set a threshold where if the back-end service fails that >> > it >> > says dead, it stays dead and needs to be manually re-added to >> > load-balancer. >> > >> > I'm trying to understand how the rise and fall settings (plus other >> > config >> > settings) can achieve this, or if there is another approach. >> > >> > Any ideas would be appreciated. >> > >> > Regards >> > >> > John >> > >> > -- >> > >> > John Clegg >> > Dash Tickets >> > http://www.dashtickets.co.nz >> > > > > > > -- > > John Clegg > Chief Technical Officier > Dash Tickets > Phone: +64 4 831 5480 x805 > http://www.dashtickets.co.nz >
