Affects also Nexus 6P
https://build.chromium.org/p/chromium.gpu.fyi/builders/Android%20Release%20%28Nexus%206P%29/builds/2110

The Labs team set up the Nexus 6P and might have more insight into its configuration.

It might be trying to bind a port that's already bound. In https://build.chromium.org/p/chromium.gpu.fyi/builders/Android%20Release%20%28Nexus%205X%29/builds/2636, it's trying to bind 50041:

(INFO) 2016-09-19 10:09:51,433 cmd_helper._ValidateAndLogCommand:153  [host]> /b/swarm_slave/w/irZl9wKz/third_party/catapult/devil/bin/deps/linux2/x86_64/forwarder_host --adb=/b/swarm_slave/w/irZl9wKz/third_party/catapult/devil/bin/deps/linux2/x86_64/bin/adb --serial-id=00ef766769cf5067 --map 50041 50041
(CRITICAL) 2016-09-19 10:09:51,439 cmd_helper.GetCmdStatusAndOutput:175  STDERR: [0919/100951:ERROR:host_forwarder_main.cc(392)] ERROR: Connection to device failed.ERROR: Existing controllers:ERROR:   40052:37278

but 50041 (0xC379) is already bound on the device:

(WARNING) 2016-09-19 10:09:51,749 android_forwarder.Create:35  Device tcp sockets in use:
(INFO) 2016-09-19 10:09:51,749 cmd_helper._ValidateAndLogCommand:153  [host]> /b/swarm_slave/w/irZl9wKz/third_party/catapult/devil/bin/deps/linux2/x86_64/bin/adb -s 00ef766769cf5067 pull /proc/net/tcp /b/swarm_slave/w/ityjedO3/tmpkjZVT3/tmp_ReadFileWithPull
(WARNING) 2016-09-19 10:09:51,950 android_forwarder.Create:39      sl  local_address rem_address   st tx_queue rx_queue tr tm->when retrnsmt   uid  timeout inode                                                     
...
(WARNING) 2016-09-19 10:09:51,956 android_forwarder.Create:39      56: 00000000:C379 00000000:0000 8A 00000000:00000000 00:00000000 00000000     0        0 22724 1 0000000000000000 100 0 0 10 0

https://build.chromium.org/p/chromium.gpu.fyi/builders/Android%20Release%20%28Nexus%205X%29/builds/2636  was run in build325-m4

Looking there I see:

chrome-bot@build325-m4:(Linux 14.04):~$ ps auxw |grep adb
chrome-+ 14602  0.0  0.0  65376  1724 ?        Ssl  13:43   0:00 /b/swarm_slave/w/irJYErNM/third_party/catapult/devil/bin/deps/linux2/x86_64/forwarder_host --adb=/b/swarm_slave/w/irJYErNM/third_party/android_tools/sdk/platform-tools/adb --serial-id=00b8096680c09d22 --map 10236 10236

However there is no swarming tasks running on this slave so it looks like some dangling adb process is to blame?

I've terminated the slave in case someone wants to take a closer look.

Labs can't fix this.  Setting it back untriaged.

Ping - please provide an update to your high priority bug. This bug is stale. Is it really P-1?

Moving Infra>Client>Android -> Infra>Client>Chrome+OS=Android