"Error: couldn't connect to server after..." when compiling TensorFlow from source on Ubuntu

[Onhge]

Hi,

I'm trying to compile TensorFlow ( https://github.com/tensorflow/tensorflow ) from source. However, when it launches "bazel clean" at the end of its configuration script, after a couple of dots and a timeout bazel displays the error "Error: couldn't connect to server after 120 seconds." I'm not getting anywhere after hours of searching and hacking, so any help would be appreciated.

After the crash, a "bazel(src)" process still exists. Connecting to it using strace shows "futex(0x7face6a539d0, FUTEX_WAIT, 22050, NULL".

Environment:

Ubuntu - 16.04.2 LTS

bazel 0.5.0 (installed as an Ubuntu package)

Python 2.7.12

$ java -version

openjdk version "9-internal"

OpenJDK Runtime Environment (build 9-internal+0-2016-04-14-195246.buildd.src)

OpenJDK 64-Bit Server VM (build 9-internal+0-2016-04-14-195246.buildd.src, mixed mode)

java.log in ~/.cache/bazel has this in it:

170606 00:10:20.992:I 1 [com.google.devtools.build.lib.analysis.BlazeVersionInfo.logVersionInfo] Blaze version info: Build label: 0.5.0

Build target: bazel-out/local-fastbuild/bin/src/main/java/com/google/devtools/build/lib/bazel/BazelServer_deploy.jar

Build time: Fri May 26 12:11:50 2017 (1495800710)

Build timestamp: 1495800710

Build timestamp as int: 1495800710

170606 00:10:21.060:I 1 [com.google.devtools.build.lib.runtime.BlazeRuntime.main] Starting Blaze server with pid pid 22048 and  and args [--max_idle_secs=10800, --connect_timeout_secs=10, --install_base=/home/user/.cache/bazel/_bazel_user/install/6d469cf67a67a9287cb9eefc2db43f7b, --install_md5=6d469cf67a67a9287cb9eefc2db43f7b, --output_base=/home/user/.cache/bazel/_bazel_user/0f792a49528a73164cd362e10f6e5a2f, --workspace_directory=/home/user/packages/src/tensorflow, --deep_execroot, --experimental_oom_more_eagerly_threshold=100, --nofatal_event_bus_exceptions, --client_debug=false, --product_name=Bazel, --option_sources=]

170606 00:10:31.481:I 14 [com.google.devtools.build.lib.profiler.AutoProfiler$LoggingElapsedTimeReceiver.log] Spent 59 milliseconds doing Idle GC

The following issues appear related, but it's unclear to me what the solution - if any - was. Is bazel incompatible with OpenJDK?

• https://github.com/tensorflow/tensorflow/issues/3694
  
• https://github.com/bazelbuild/bazel/issues/1636
  
• https://github.com/bazelbuild/bazel/issues/1638
  

Thanks!

[Steren Giannini]

Hi,

This is more a TensorFlow related question, please reach the TensorFlow support channels for an answer, I suggest https://stackoverflow.com/questions/tagged/tensorflow 

As far as I know, Bazel is compatible with Open JDK, on some platform, we are even bundling Bazel with OpenJDK to ease installation

--
You received this message because you are subscribed to the Google Groups "bazel-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bazel-discus...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bazel-discuss/4d4bacb9-008b-4465-8b0b-b33f7d7753de%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.