NEO installs but fails to start (path issue?)

NEO has successfully installed with no errors. But the service fails to start:

/opt/neo/neoservice start

checking ha-mode : … [ OK ]

Checking Localhost resolving … [ OK ]

Checking Hostname resolving … [ OK ]

Starting neo-access-credentials: [FAILED]

Starting neo-controller: [FAILED]

Starting neo-device-manager: [FAILED]

Starting neo-eth-discovery: [FAILED]

Starting neo-host-manager: [FAILED]

Starting neo-ib: [FAILED]

Starting neo-ip-discovery: [FAILED]

Starting neo-monitor: [FAILED]

Starting neo-performance: [FAILED]

Starting neo-provisioning: [FAILED]

Starting neo-solution: [FAILED]

Starting neo-telemetry: [FAILED]

Starting neo-virtualization: [FAILED]

The first service fails to start:

systemctl start neo-access-credentials

Job for neo-access-credentials.service failed because the control process exited with error code. See “systemctl status neo-access-credentials.service” and “journalctl -xe” for details.

systemctl status neo-access-credentials.service

neo-access-credentials.service - LSB: Start the NEO-ACCESS-CREDENTIALS

Loaded: loaded (/etc/rc.d/init.d/neo-access-credentials; bad; vendor preset: disabled)

Active: failed (Result: exit-code) since Thu 2019-04-11 16:40:19 PDT; 8s ago

Docs: man:systemd-sysv-generator(8)

Process: 40237 ExecStart=/etc/rc.d/init.d/neo-access-credentials start (code=exited, status=1/FAILURE)

Apr 11 16:40:12 krgs3.eth systemd[1]: Starting LSB: Start the NEO-ACCESS-CREDENTIALS…

Apr 11 16:40:13 krgs3.eth python[40255]: detected unhandled Python exception in ‘/opt/neo/providers/ac/bin/ac/ac_service.pyo’

Apr 11 16:40:13 krgs3.eth python[40255]: can’t communicate with ABRT daemon, is it running? [Errno 2] No such file or directory

Apr 11 16:40:19 krgs3.eth neo-access-credentials[40237]: Starting neo-access-credentials: [FAILED]

Apr 11 16:40:19 krgs3.eth systemd[1]: neo-access-credentials.service: control process exited, code=exited status=1

Apr 11 16:40:19 krgs3.eth systemd[1]: Failed to start LSB: Start the NEO-ACCESS-CREDENTIALS.

Apr 11 16:40:19 krgs3.eth systemd[1]: Unit neo-access-credentials.service entered failed state.

Apr 11 16:40:19 krgs3.eth systemd[1]: neo-access-credentials.service failed.

Is this a path problem? I can make some progress by manipulating the PYTHONPATH, but eventually I get stuck:

[root@krgs3 ac]# PYTHONPATH=/opt/neo/providers/ac/bin:/opt/neo/common/bin:/opt/neo/providers/common/bin:/opt/neo/common/bin/netservices/common/infra python ac_service.pyo

Traceback (most recent call last):

File “/opt/neo/providers/ac/bin/ac/ac_service.py”, line 24, in

File “/opt/neo/providers/common/bin/providers/common/netservice.py”, line 45, in

File “/opt/neo/common/bin/netservices/common/infra/web/wsgi_srv.py”, line 17, in

ImportError: No module named web

[root@krgs3 ac]# PYTHONPATH=/opt/neo/providers/ac/bin python ac_service.pyo

Traceback (most recent call last):

File “/opt/neo/providers/ac/bin/ac/ac_service.py”, line 17, in

File “/opt/neo/providers/ac/bin/ac/ac_discovery.py”, line 19, in

ImportError: No module named netservices.common.infra.tools.fswatcher

How is this supposed to work?

Thanks.

Meant to show loading the .pyo without any PYTHONPATH:

[root@krgs3 ac]# python ac_service.pyo

Traceback (most recent call last):

File “/opt/neo/providers/ac/bin/ac/ac_service.py”, line 16, in

ImportError: No module named ac

I can hack PYTHONPATH to get some of the modules to load, but I get stuck on “web”. Surely this isn’t how it’s supposed to work, just trying to debug it.

Hello Joseph,

I hope all is good.

What OS flavor and version are you using?

What NEO version are you using?

Could you try:

  • Check if any packages were pulled in from repositories during NEO install.
  • Uninstall NEO.
  • Disable repositories.
  • Reinstall NEO.

Many thanks -

~Steve

Hello Joseph,

NEO: 2.3.1 is latest version, OS RedHat/CentOS 7.3, 7.4, 7.5

Could you take a look at the Release Notes to see if your OS is supported? https://mymellanox.force.com/support/SupportProductItem?id=a2v50000000XrBHAA0

Many thanks,

~Steve

Thanks – yes, downgrading to CentOS 7.4 resulted in a working installation. Do note that the requirements for server installation are inconsistent between Release Notes, Quick Start Guide, and User Manual for v2.3.1