Very rarely (?) swaps results across hosts, when pinging multiple

[romanrm]

Hello,

I run oping giving it a list of about 15 IPs to ping, and the count to ping is 3.

Just now I had a really weird issue, where it swapped the ping time result of two hosts. The output was otherwise correct.

Grepping the log for "Host A" (actually ULA IPv6) returned:

56 bytes from (Host A): icmp_seq=1 ttl=64 time=81.87 ms
56 bytes from (Host A): icmp_seq=2 ttl=64 time=83.55 ms
56 bytes from (Host A): icmp_seq=3 ttl=64 time=81.20 ms

But there is definitely no way that "Host A" would ping at 81ms, due to the geographic distance.

At the same time, there was "Host B" in the same log which pinged at 311ms, which would be exactly the expected value for "Host A". Whereas normal ping times for "Host B" would be around those 81ms.
There is definitely no mistake with IPs used, and no routing or IP assignment mistakes.
It would appear that these two hosts got their ping times swapped in the output log.

Unfortunately I don't know how to reproduce. And neither before, nor the next time (this is done in cron every 15 minutes) the issue has not repeated.

I use oping 1.10.0 from the Debian package version 1.10.0-2.1+b1 on amd64.