1 root root 22 Nov 11 09:38 python -> /usr/libexec/no-python 1 root root 20 Nov 11 09:32 ifup -> /usr/libexec/nm-ifup 1 root root 22 Nov 11 09:32 ifdown -> /usr/libexec/nm-ifdown 1 root root 39 Nov 11 09:32 cifs-idmap-plugin -> /usr/lib64/cifs-utils/cifs_idmap_sss.so There is nothing in alternatives that is configured for /usr/local in pip. Yum remove and install python3 (I don't think removing python3 is a good idea just wanted to test it - it did not fix pip). Update-alternatives because I noticed python3 is in 'manual' now compared to 'auto' on a fresh rhel8 install. Ran python3 -m pip install pip=9.0.3 but this is just 9.0.3 on /usr/local/ instead of /usr/lib. I've been setting up fresh VMs the last 3 hours and can't find a solution.

I've since used python3 -m pip install pip -upgrade -user and/or updated the ansible with become_user=appuser.Ĭouple things I've tried that did not work.

The default is using pip 9.0.3 from /usr/lib/python3.6/site-packages (python 3.6)Īfter the mistake pip is 20.3.3 from /usr/local/lib/python3.6/site-packages/pip (python 3.6)

I've gotten that done with -user argument for new deploys but I need to fix existing 4 servers.ĭoing some reading around it was pretty obviously this was a bad idea!! Is there a way to get the system-wide pip state back to normal? I required a newer version of pip 20.x for a module for an application user only. So it took the become for the playbook with root.

On some RHEL8.3 box's I've accidentally upgraded root's pip version with an ansible playbook because I did not set become_user=appuser in the pip module.

Pip3 upgrade