Tests failures: System.Net.NameResolution.Tests / DnsObsolete* & Dns_GetHostEntry*

Type of failures

Affected tests:

• System.Net.NameResolution.Tests.GetHostByNameTest (both always fail together)
  • DnsObsoleteGetHostByName_EmptyString_ReturnsHostName
  • DnsObsoleteBeginEndGetHostByName_EmptyString_ReturnsHostName
• System.Net.NameResolution.Tests.GetHostEntryTest (both always fail together)
  • Dns_GetHostEntry_HostString_Ok
  • Dns_GetHostEntryAsync_HostString_Ok

Failure text:

• OSX: Device not configured
• Linux: No such device or address

System.Net.Internals.SocketExceptionFactory+ExtendedSocketException : Device not configured
OR
System.Net.Internals.SocketExceptionFactory+ExtendedSocketException : No such device or address

at System.Net.Dns.InternalGetHostByName(String hostName, Boolean includeIPv6) in /Users/buildagent/agent/_work/34/s/corefx/src/System.Net.NameResolution/src/System/Net/DNS.cs:line 46
at System.Net.Dns.GetHostByName(String hostName) in /Users/buildagent/agent/_work/34/s/corefx/src/System.Net.NameResolution/src/System/Net/DNS.cs:line 41
at System.Net.NameResolution.Tests.GetHostByNameTest.DnsObsoleteGetHostByName_EmptyString_ReturnsHostName() in /Users/buildagent/agent/_work/34/s/corefx/src/System.Net.NameResolution/tests/FunctionalTests/GetHostByNameTest.cs:line 108

Note:

• Dns_GetHostEntryAsync_HostString_Ok fails almost always with same error twice, wrapped in System.AggregateException.

Older failure on OSX (prior to 2017/9/15)

Assert.Contains() Failure
Not found: DCI-Mac-Build-068.local
In value:  dci-mac-build-068.local
at System.Net.NameResolution.Tests.GetHostByNameTest.DnsObsoleteGetHostByName_EmptyString_ReturnsHostName() in /Users/buildagent/agent/_work/30/s/corefx/src/System.Net.NameResolution/tests/FunctionalTests/GetHostByNameTest.cs:line 108

History of failures

Each test couple always fails or passes - maybe bad machine setup?

Day	Build	OS	Test
9/6	20170906.01	OSX10.12	DnsObsolete*
9/10	20170910.01	Ubuntu14.04	DnsObsolete* & Dns_GetHostEntry*
9/13	20170913.02	OSX10.12	DnsObsolete*
9/15	20170915.01	OSX10.12	DnsObsolete*
9/20	20170920.03	Debian90	DnsObsolete* & Dns_GetHostEntry*
9/30	20170930.01	OSX10.12	DnsObsolete* & Dns_GetHostEntry*
10/2	20171002.01	Debian87	DnsObsolete* & Dns_GetHostEntry*
10/21	20171021.01	Debian87	DnsObsolete* & Dns_GetHostEntry*
10/26	20171026.01	Suse42.2	DnsObsolete* & Dns_GetHostEntry*
11/13	20171113.03	OSX10.12	DnsObsolete* & Dns_GetHostEntry*
11/16	20171116.51	SLES12	DnsObsolete* & Dns_GetHostEntry*
11/26	20171126.02	Suse42.2	DnsObsolete* & Dns_GetHostEntry*

Similar failures to dotnet/corefx#20245, but at different times (probably runs on different worker machine).

Runfo Tracking Issue: dnsobsoletegethostbyname_emptystring_returnshostname

Build	Definition	Kind	Run Name

Build Result Summary

Day Hit Count	Week Hit Count	Month Hit Count
0	0	0

[Sunny-pu]

[EDIT] Test failure included in history in top post.

[karelz]

Updated test failure history in top post.

[Sunny-pu]

failed again:
https://mc.dot.net/#/product/netcore/master/source/official~2Fcorefx~2Fmaster~2F/type/test~2Ffunctional~2Fcli~2F/build/20180419.01/workItem/System.Net.NameResolution.Functional.Tests

[benaadams]

Seen in dotnet/corefx#35267

Debian.8.Amd64.Open-x64-Release

Message :
System.Net.Internals.SocketExceptionFactory+ExtendedSocketException : Name or service not known
Stack Trace :
   at System.Net.Dns.InternalGetHostByName(String hostName) in /__w/1/s/src/System.Net.NameResolution/src/System/Net/DNS.cs:line 68
   at System.Net.Dns.GetHostByName(String hostName) in /__w/1/s/src/System.Net.NameResolution/src/System/Net/DNS.cs:line 41
   at System.Net.NameResolution.Tests.GetHostByNameTest.DnsObsoleteGetHostByName_EmptyString_ReturnsHostName() in /__w/1/s/src/System.Net.NameResolution/tests/FunctionalTests/GetHostByNameTest.cs:line 108

[AriNuer]

Same test failed with similar exception:

Failed test:
System.Net.NameResolution.Tests.GetHostByNameTest/DnsObsoleteBeginEndGetHostByName_EmptyString_ReturnsHostName

Message:

System.Net.Internals.SocketExceptionFactory+ExtendedSocketException : Name or service not known

Stack Trace :

   at System.Net.Dns.InternalGetHostByName(String hostName) in /_/src/System.Net.NameResolution/src/System/Net/DNS.cs:line 68
   at System.Net.Dns.ResolveCallback(Object context) in /_/src/System.Net.NameResolution/src/System/Net/DNS.cs:line 218
--- End of stack trace from previous location where exception was thrown ---
   at System.Runtime.ExceptionServices.ExceptionDispatchInfo.Throw(Exception source) in /_/src/System.Private.CoreLib/shared/System/Runtime/ExceptionServices/ExceptionDispatchInfo.cs:line 69
   at System.Net.Dns.HostResolutionEndHelper(IAsyncResult asyncResult) in /_/src/System.Net.NameResolution/src/System/Net/DNS.cs:line 361
   at System.Net.Dns.EndGetHostByName(IAsyncResult asyncResult) in /_/src/System.Net.NameResolution/src/System/Net/DNS.cs:line 383
   at System.Net.NameResolution.Tests.GetHostByNameTest.DnsObsoleteBeginEndGetHostByName_EmptyString_ReturnsHostName() in /_/src/System.Net.NameResolution/tests/FunctionalTests/GetHostByNameTest.cs:line 119

Build: -20190617.81(Master)
Failing configurations:

• RedHat.6.Amd64.Open-x64-Release
• Debian.9.Amd64.Open-x64-Release

Details:
https://mc.dot.net/#/user/dotnet-bot/pr~2Fdotnet~2Fcorefx~2Frefs~2Fheads~2Fmaster/test~2Ffunctional~2Fcli~2Finnerloop~2F/20190617.81/workItem/System.Net.NameResolution.Functional.Tests/analysis/xunit/System.Net.NameResolution.Tests.GetHostByNameTest~2FDnsObsoleteBeginEndGetHostByName_EmptyString_ReturnsHostName

[karelz]

Triage: We should know why DNS resolution does not work sometimes -- we know it is not cached on Linux/Mac, but why does it happen so often in CI? Is it network flakiness? Or busy machine?

Plan: Add production diagnostics into DNS, then make runs with those changes to catch it. It will be useful also for production diagnostic by customers.