This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Fwd: Increasing resolver timeout


For example we have the following bug:

$ time ping -c1 "google.com""
0% packet loss
0m5,064s

$ time (ip=$(getent ahostsv4 "google.es" | awk '{ print $1 }' |
head -n1); ping -c1 "${ip}")
0% packet loss
0m0,042s

The extra time on the first case it's caused by a router mistakenly not returning the IPv6 DNS reply, so the command just sits till it timeouts.

The point is this, the bug has been there for a long time unnoticed simply because with 5s timeout the delay simply felt within normal. It wasn't till I started programing server stuff where I noticed that retrieving the data just took so long.

So probably we would be better having a longer timeout, like about 10s. So if there is any network anomaly it would be obvious to any person operating it.

Alberto <https://es20490446e.wordpress.com>


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]