#12685 The cloud-noc-os01 machine seems to hang on SSH
Closed: Fixed with Explanation by zlopez. Opened by praiskup.

Making us unable to reboot hypervisors according to:
https://docs.fedoraproject.org/en-US/infra/sysadmin_guide/hardware_troubleshooting_power/

https://matrix.to/#/!ypZEUFVkgStvzXNAnr:matrix.org/$8J-M_cEftHJJc0irxG5hSkBg_dFz4BhiNOaXEiSrlDo?via=fedoraproject.org&via=fedora.im&via=matrix.org


Metadata Update from @zlopez:
- Issue priority set to: Waiting on Assignee (was: Needs Review)
- Issue tagged with: Needs investigation, medium-gain, ops

Ah, docs are just out of date there.

Should be:

noc-cc01.rdu-cc.fedoraproject.org

It made little sense to call something a 'cloud-noc-os01' when it was not for cloud, not for openshift and only a noc.

Please use the above host and we should update the docs.

I suppose we could have it resolve to the same ip in dns?

Metadata Update from @praiskup:
- Issue assigned to praiskup

I was about to fix this:
https://pagure.io/infra-docs-fpo/pull-request/384

But I still get

$ ssh -i ~/.ssh/fedora -ND 9999 noc-cc01.rdu-cc.fedoraproject.org
praiskup@noc-cc01.rdu-cc.fedoraproject.org: Permission denied (publickey,gssapi-keyex,gssapi-with-mic).

Metadata Update from @praiskup:
- Assignee reset

Metadata Update from @kevin:
- Issue assigned to kevin

Try now? sssd seems to have disconnected there... possibly it wasn't restarted right after switching ipa servers.

Actually, doesn't work - but I might be doing an unsupported thing (that used to work before though):

$ ssh -i ~/.ssh/fedora -ND 9999 noc-cc01.rdu-cc.fedoraproject.org
Connection closed by 8.43.85.50 port 22

sshuttle asks me for sudo password, and that's what I don't want to do.

Here's what I do:

On my local laptop:

sshuttle 172.23.1.0/24 -r noc-cc01.rdu-cc.fedoraproject.org -v

yes, it will ask for sudo, but only for local laptop (to setup the nftables/iptables rules)

Then you can just access 172.23.1.x ip's directly on your laptop while the tunnel is up.

The ssh works for me too here... are you able to just ssh directly normally to noc-cc01? No need to gateway via bastion, just direct?

Oh, I also noticed that host had a 100% full /
:(

Fixed that... perhaps some of these issues are related to that. ;( We should both make it larger and clean up our tftpboot image files.

It works now! Thank you.

Metadata Update from @zlopez:
- Issue close_status updated to: Fixed with Explanation
- Issue status updated to: Closed (was: Open)

This stay opened: https://pagure.io/infra-docs-fpo/pull-request/384

Metadata
Related Pull Requests