Oracle Linux 8 ARM64 - Fatal Python error: init_fs_encoding

[Matty-uk]

Hi

I've just done a source build to update from ncpa 3.0.0 to 3.1.0. Everything appeared to go well both with the build from source and the install/upgrade. However, when I do a systemctl status I get the following:

systemctl status ncpa
● ncpa.service - NCPA
Loaded: loaded (/usr/lib/systemd/system/ncpa.service; enabled; vendor preset: disabled)
Active: failed (Result: exit-code) since Fri 2018-06-22 12:12:05 BST; 6 years 0 months ago
Docs: https://www.nagios.org/ncpa
Process: 660 ExecStart=/usr/local/ncpa/ncpa -n (code=exited, status=1/FAILURE)
Main PID: 660 (code=exited, status=1/FAILURE)

Jun 22 12:12:05 nagiospd02l-hu ncpa[660]: sys.path = [
Jun 22 12:12:05 nagiospd02l-hu ncpa[660]: '/usr/local/ncpa/lib',
Jun 22 12:12:05 nagiospd02l-hu ncpa[660]: ]
Jun 22 12:12:05 nagiospd02l-hu ncpa[660]: Fatal Python error: init_fs_encoding: failed to get the Python codec of the filesystem encoding
Jun 22 12:12:05 nagiospd02l-hu ncpa[660]: Python runtime state: core initialized
Jun 22 12:12:05 nagiospd02l-hu ncpa[660]: ModuleNotFoundError: No module named 'encodings'
Jun 22 12:12:05 nagiospd02l-hu ncpa[660]: Current thread 0x0000fffc72f74630 (most recent call first):
Jun 22 12:12:05 nagiospd02l-hu ncpa[660]:
Jun 22 12:12:05 nagiospd02l-hu systemd[1]: ncpa.service: Main process exited, code=exited, status=1/FAILURE
Jun 22 12:12:05 nagiospd02l-hu systemd[1]: ncpa.service: Failed with result 'exit-code'.

Rolling back to version 3.0.0 that I build a while back works fine.

Built on a clean/fresh install of Oracle Linux 8.10 with all patches applied.

Any suggestion on how to get this working appreciated.

[gittethis]

Not sure about the prereq to make a linux build but i’ guessing there is at least one (encodings) import missing(not downloaded into the agent files tree) during build so it will yield NotFound when preparing the python env/run at RUNTIME.(Most likely a line/word to be added)On 26 Jun 2024, at 15:25, Matty-uk ***@***.***> wrote: Hi I've just done a source build to update from ncpa 3.0.0 to 3.1.0. Everything appeared to go well both with the build from source and the install/upgrade. However, when I do a systemctl status I get the following: systemctl status ncpa ● ncpa.service - NCPA Loaded: loaded (/usr/lib/systemd/system/ncpa.service; enabled; vendor preset: disabled) Active: failed (Result: exit-code) since Fri 2018-06-22 12:12:05 BST; 6 years 0 months ago Docs: https://www.nagios.org/ncpa Process: 660 ExecStart=/usr/local/ncpa/ncpa -n (code=exited, status=1/FAILURE) Main PID: 660 (code=exited, status=1/FAILURE) Jun 22 12:12:05 nagiospd02l-hu ncpa[660]: sys.path = [ Jun 22 12:12:05 nagiospd02l-hu ncpa[660]: '/usr/local/ncpa/lib', Jun 22 12:12:05 nagiospd02l-hu ncpa[660]: ] Jun 22 12:12:05 nagiospd02l-hu ncpa[660]: Fatal Python error: init_fs_encoding: failed to get the Python codec of the filesystem encoding Jun 22 12:12:05 nagiospd02l-hu ncpa[660]: Python runtime state: core initialized Jun 22 12:12:05 nagiospd02l-hu ncpa[660]: ModuleNotFoundError: No module named 'encodings' Jun 22 12:12:05 nagiospd02l-hu ncpa[660]: Current thread 0x0000fffc72f74630 (most recent call first): Jun 22 12:12:05 nagiospd02l-hu ncpa[660]: Jun 22 12:12:05 nagiospd02l-hu systemd[1]: ncpa.service: Main process exited, code=exited, status=1/FAILURE Jun 22 12:12:05 nagiospd02l-hu systemd[1]: ncpa.service: Failed with result 'exit-code'. Rolling back to version 3.0.0 that I build a while back works fine. Built on a clean/fresh install of Oracle Linux 8.10 with all patches applied. Any suggestion on how to get this working appreciated. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[Matty-uk]

I'll see if I can track down what's missing and provide the details.

[ne-bbahn]

After a bit of research, it seems that this problem usually occurs when the PYTHONHOME or PYTHONPATH variables are set wrongly. I had this issue on one CentOS 9 machine, but when I tried building the same branch on another Cent9 machine, it worked without issue. This definitely seems to be a build issue.

[ne-bbahn]

Upon further testing, this doesn't seem to be the cause here, though. I will continue to investigate why this might be happening.

[ne-bbahn]

I checked the above metrics against my good build and it seems that sys.path in the working version includes /usr/local/ncpa/lib/library.zip in addition to /usr/local/ncpa/lib, but all the other metrics above are identical.

[ne-bbahn]

Looks like this has happened before with cx_freeze marcelotduarte/cx_Freeze#835

It seems that the issue arises from cx_Freeze 7, so there is likely some process that needs to be updated.