0

I have tried to follow the quickstart of Google TPU setup as the following link.

https://cloud.google.com/tpu/docs/quickstart

When I run "ctpu up" in the google cloud shell, it shows the following error.

*2018/08/05 09:37:08 Created TPU xxx-xxx!
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x18 pc=0x6677b3]
goroutine 1 [running]:
github.com/tensorflow/tpu/tools/ctpu/ctrl.(*ResourceManagementCP).IsProjectInGoogleOrg(0xc4200511b0, 0x773340, 0xc42035
c640, 0xc42000c798)
        /tmp/ctpu-release/src/github.com/tensorflow/tpu/tools/ctpu/ctrl/resourcemgmt.go:119 +0xe3
github.com/tensorflow/tpu/tools/ctpu/commands.(*upCmd).Execute(0xc4200795e0, 0x773380, 0xc4200160d0, 0xc4200568a0, 0x0,
 0x0, 0x0, 0x6e05a0)
        /tmp/ctpu-release/src/github.com/tensorflow/tpu/tools/ctpu/commands/up.go:449 +0x2c3
github.com/google/subcommands.(*Commander).Execute(0xc420070000, 0x773380, 0xc4200160d0, 0x0, 0x0, 0x0, 0x5)
        /tmp/ctpu-release/src/github.com/google/subcommands/subcommands.go:141 +0x29f
github.com/google/subcommands.Execute(0x773380, 0xc4200160d0, 0x0, 0x0, 0x0, 0xc420052700)
        /tmp/ctpu-release/src/github.com/google/subcommands/subcommands.go:385 +0x5f
main.main()
        /tmp/ctpu-release/src/github.com/tensorflow/tpu/tools/ctpu/main.go:87 +0xd5e*

Even I tried to ssh to the new created VM, I cannot run those sample code.

This is the first step of the setup in the above tutorial with the official google environment but it was also fail.

Would anyone can help to advise if there is any missing step?

Thanks a lot

Armin_SC
  • 2,130
  • 11
  • 15
Thomas Lau
  • 121
  • 1
  • I believe `TPU` is not available while you are running in trial mode. You may have to upgrade your project to paid one. – Asdfg Aug 06 '18 at 17:10

1 Answers1

0

This error message seems to be generated by a reference piece of code used by the TPU components to determine if the project is part of the Google organization.

Based on this, a github issue has been created to report this unexpected behavior where it is explained the following workarounds to avoid this error message:

Armin_SC
  • 2,130
  • 11
  • 15