ping output shows truncated entry from not pinged address
I noticed following strange output when pinging a node: [prompt] # ping x.x.x.x PING x.x.x.x (x.x.x.x) 56 (84) bytes of data. 64 bytes from x.x.x.x: icmp_seq=1 ttl=61 time=xx ms 64 bytes from x.x.x.x: icmp_seq=2 ttl=61 time=xx ms 64 bytes from x.x.x.x: icmp_seq=3 ttl=61 time=xx ms 64 bytes from x.x.x.x: icmp_seq=4 ttl=61 time=xx ms 64 bytes from x.x.x.x: icmp_seq=5 ttl=61 time=xx ms ... 32 bytes from y.y.y.y: icmp_seq=26978 ttl=61 (truncated) 64 bytes from x.x.x.x: icmp_seq=54 ttl=61 time=xx ms 64 bytes from x.x.x.x: icmp_seq=55 ttl=61 time=xx ms When stopping the ping I see in the statistics one more packet received than transmitted. Where does this comes from? Has this any special meaning? Is this indicating any strange behavior or misconfiuration in the network? Thanks for any helpful ideas or information to further investigate this! Regards, Stefan 🙂
Health monitor to mark pool member up when a ping FAILS (inverse ping health check)
Hello, I am attempting to figure out the best way to create a health monitor that is an inverse of the ICMP monitor-when ping to an IP fails, I want the pool member to be marked UP. When ping to an IP succeeds, I want the pool member to be marked DOWN. My first attempt at this was using an external monitor with the below script; that does not seem to be working. #!/bin/sh ping -c 5 [IP here] > /dev/null && exit || echo "down" exit $? I am running 11.5.3 HF2. Does anyone know if the best way to achieve this beahvior, or why my script isn't working? It looks like it outputs nothing to stdout (should cause the monitor to fail) unless the ping fails, at which point it outputs something (should cause the monitor to pass)
Using Selective ACK on Virtual Address
We have a 2 tier system whereby Tier 1 needs to know quickly if Tier 2 services have stopped, be that VS or the pools or the box itself. To that end we've setup an ICMP monitor at tier 1 to ping the VS at Tier 2 and we've configured the ICMP option within Virtual Address to be Selective (ie don't respond if VS is red). This works well most of the time. However every now and again we'll see that a service has been marked down at Tier 1 and the ICMP monitor is the one saying it had marked the service down. You go to Tier 2 and the VS is green (it has been down but come back up) and a TCPDUMP shows that the ICMP request is getting to Tier 2 but Tier 2 is NOT sending an ICMP response. Typically the only fix is to reboot the Tier 2 box - which is not ideal in a production environment. I have opened a case with F5 for this but wondered if anybody else had come across something similar etc.
Ping received route from BGP and Kernel fails
Have difficulties in pinning an address which I receive as part of BGP adverrtisement. The same also for the address which comes through an address which I have advertised to the Kernel. The ping gives : connect: Network is unreachable Altho I have the route in my routing table. Any thoughts please
