schweikert
Inconsistent behavior with -g option
fping shows the same (alive) host as being both unreachable and alive depending on the CIDR subnet given to the -g option. Here is what I see:
# fping -v
fping: Version 4.2
fping: comments to [email protected]
# fping -qug 192.168.32.0/23 | grep 254
192.168.32.254
192.168.33.254
# fping -qag 192.168.33.240/28
192.168.33.241
192.168.33.242
192.168.33.243
192.168.33.244
192.168.33.245
192.168.33.246
192.168.33.247
192.168.33.248
192.168.33.249
192.168.33.250
192.168.33.251
192.168.33.252
192.168.33.253
192.168.33.254
Why does 192.168.33.254, which is definitely alive, appear in the output of both commands? It should appear in the alive output only and not in the unreachable output.
Testing against fping 5.1 I don't see this issue, below I have 192.168.1.254 as reachable and 192.168.0.254 as unreachable.
docker run fping -v
fping: Version 5.1
# 1.254 is alive
docker run fping 192.168.1.254
192.168.1.254 is alive
$ docker run fping -qug 192.168.1.0/23 | grep 254
192.168.0.254
$ docker run fping -qag 192.168.1.240/28
192.168.1.254
Also tested against 4.2 (using v4.2 tag) also don't see the issue.
$ docker run fping:4.2 -v
fping: Version 4.2
fping: comments to [email protected]
$ docker run fping:4.2 -qug 192.168.1.0/23 | grep 254
192.168.0.254
docker run fping:4.2 -qag 192.168.1.240/28
192.168.1.254
Since I can't reproduce, what does tcpdump show you when you run the command, do you get a response for 192.168.33.254 when you do the /23 mask, since that's testing a lot more addresses at once than /28. Doing a tcpdump for those ranges in testing captured 7k packets for /23 vs 213 packets for /28.
Perhaps the system with IP address 192.168.33.254 uses ICMP rate limiting, or has an unstable connection, or is sometimes overloaded, and thus sometimes appears unreachable, sometimes reachable?
I do not expect this to be an fping issue.
