On 10/10/16 2:17 AM, Lukas Pirl wrote: > On 10/09/2016 11:15 PM, Brian Warner wrote as excerpted: >> Thanks.. could you check that those two slaves have "127.0.0.1" bound to >> the loopback interface? One of the failing tests suggests that they do >> not, and I think the other tests depend upon being able to use 127.0.0.1 >> to connect to themselves. > > They do > > [tahoe@tahoe-buildbot-centos7 root]$ ip addr show lo > 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1 > link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 > inet 127.0.0.1/8 scope host lo > valid_lft forever preferred_lft forever > inet6 ::1/128 scope host
Hmm. What's the full output of "/bin/ip addr" on that box? Tahoe tries a bunch of different commands to find the addresses (starting with /bin/ip, then a bunch of ifconfig flavors). It applies a regexp to the output of each (see src/allmydata/util/iputil.py line 164 "_unix_commands"). I'm wondering if there's something unusual about this host's output, and the regexp isn't correctly matching it. allmydata.test.test_iputil.ListAddresses.test_list_async is the failure I'm looking at (using the "fix the simplest thing first in the hopes that it will help the complicated things" debugging philosophy). It's reporting 10.10.1.14, but not 127.0.0.1. The test_list_async() function it calls combines the "/bin/ip" / "ifconfig" output with a routine that prepares a UDP socket for talking with one of the root DNS servers, then asking what interface that socket would have used. If ip/ifconfig is broken entirely, it might still be getting 10.10.1.14 from the UDP socket routine. thanks, -Brian _______________________________________________ tahoe-dev mailing list tahoe-dev@tahoe-lafs.org https://tahoe-lafs.org/cgi-bin/mailman/listinfo/tahoe-dev
