Repository navigation
test: DNS for bare domain fails with fritz.box search domain #9799
Copy link
Copy link
Closed
Labels
dnsIssues and PRs related to the dns subsystem.Issues and PRs related to the dns subsystem.good first issueIssues that are suitable for first-time contributors.Issues that are suitable for first-time contributors.testIssues and PRs related to Node.js core tests and test infrastructure.Issues and PRs related to Node.js core tests and test infrastructure.
Description
- Version: master
- Platform: Linux localhost.localdomain 4.8.8-300.fc25.x86_64 deps: update openssl to 1.0.1j #1 SMP Tue Nov 15 18:10:06 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
- Subsystem: dns
Discovered while on a LAN with a fritz.box search domain.
$ sudo sh -c 'echo "search fritz.box" >> /etc/resolv.conf'
$ out/Release/node test/parallel/test-net-better-error-messages-port-hostname.js
assert.js:85
throw new assert.AssertionError({
^
AssertionError: 'Unknown system error -74' == 'ENOTFOUND'
at Socket.<anonymous> (/home/bnoordhuis/src/v1.x/test/parallel/test-net-better-error-messages-port-hostname.js:11:10)
at Socket.<anonymous> (/home/bnoordhuis/src/v1.x/test/common.js:422:15)
at emitOne (events.js:96:13)
at Socket.emit (events.js:188:7)
at connectErrorNT (net.js:1025:8)
at _combinedTickCallback (internal/process/next_tick.js:74:11)
at process._tickCallback (internal/process/next_tick.js:98:9)
$ strace -s 512 -fe socket,connect,recvfrom,sendmmsg out/Release/node test/parallel/test-net-better-error-messages-port-hostname.js
[...]
[pid 25161] socket(PF_INET, SOCK_DGRAM|SOCK_NONBLOCK, IPPROTO_IP) = 12
[pid 25161] connect(12, {sa_family=AF_INET, sin_port=htons(53), sin_addr=inet_addr("8.8.8.8")}, 16) = 0
[pid 25161] sendmmsg(12, {{{msg_name(0)=NULL, msg_iov(1)=[{"\3102\1\0\0\1\0\0\0\0\0\0\3***\5fritz\3box\0\0\1\0\1", 31}], msg_controllen=0, msg_flags=MSG_DONTWAIT|MSG_EOR|MSG_SYN|MSG_RST|MSG_CMSG_CLOEXEC|0x1ed80010}, 31}, {{msg_name(0)=NULL, msg_iov(1)=[{"\322 \1\0\0\1\0\0\0\0\0\0\3***\5fritz\3box\0\0\34\0\1", 31}], msg_controllen=0, msg_flags=MSG_OOB|MSG_PEEK|MSG_CTRUNC|MSG_WAITALL|MSG_FIN|MSG_SYN|MSG_CONFIRM|MSG_WAITFORONE|MSG_FASTOPEN|0x85b40000}, 31}}, 2, MSG_NOSIGNAL) = 2
[pid 25161] recvfrom(12, "\3102\201\200\0\1\0\1\0\0\0\0\3***\5fritz\3box\0\0\1\0\1\300\f\0\1\0\1\0\0\t\262\0\4\177\00055", 2048, 0, {sa_family=AF_INET, sin_port=htons(53), sin_addr=inet_addr("8.8.8.8")}, [16]) = 47
[pid 25161] recvfrom(12, "\322 \201\200\0\1\0\0\0\1\0\0\3***\5fritz\3box\0\0\34\0\1\300\26\0\6\0\1\0\0\3\203\0005\1a\3nic\300\26\7support\variservices\3com\0X\37\325\341\0\0\7\10\0\0\1,\0\33\257\200\0\0\7\10", 65536, 0, {sa_family=AF_INET, sin_port=htons(53), sin_addr=inet_addr("8.8.8.8")}, [16]) = 96
Error -74 is EBADMSG. Changing the search domain to something more benign (e.g. domain.local) makes the test pass again. There are more tests with the same issue.
Tests in test/parallel and test/sequential should not need internet access so perhaps it's best to move affected tests to test/internet but that isn't run regularly. Either that or they should be modified to not make actual DNS queries.
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
dnsIssues and PRs related to the dns subsystem.Issues and PRs related to the dns subsystem.good first issueIssues that are suitable for first-time contributors.Issues that are suitable for first-time contributors.testIssues and PRs related to Node.js core tests and test infrastructure.Issues and PRs related to Node.js core tests and test infrastructure.
Activity
thelostone-mc commentedon Jun 29, 2017
@bnoordhuis / @thefourtheye Is this still a valid use case ?
I was able to run the tests in parallel, sequential without an active internet connection ?
bnoordhuis commentedon Jun 29, 2017
@adityaanandmc I think this is still relevant. It's not so much about no internet connection as it is about the search domain making non-internet tests fail.
Marsup commentedon Nov 6, 2017
I'm not really sure but I think #14781 fixed it, or at least I can't make it fail with your instructions anymore. Would you say it is fixed ?
bnoordhuis commentedon Nov 6, 2017
@Marsup Yes, looks like it. Thanks, I'll close this out.