cumulus vx-3.2.1 NCLU

[Sipart]

Hi - the cumulus vx 3.2.1 is not listed as a supported image but seems to basically work when editing the seperate files in a Linux style.

Hardware : CPU - Dual 2.66ghz Quad core Xeons with 24gb ram, 1x300GB and 2x146GB 10K SAS drives (one volume), 1 network interface

EVE-NG Running on - Bare Metal Linux (Ubunut 16.10)

VT-X - Enabled

Lots of other images working fine (vMX, vSRX, Cisco, Win7 etc)

But I do have some issues when trying to use the NCLU on this image (https : // docs.cumulusnetworks.com/display/DOCS/Network+Command+Line+Utility) which is included in 3.2.1.

But as its not a supported image I wonderderd if anyone else had got 3.2.1 NCLU working and if so how?

I tried before and after doing a apt-get update on the cumulus VX.

When I try to save or commit a config I get the following:

"cumulus@sw01-nni-1:~$ net show configuration

ERROR: net could not connect to netd

Try starting netd with:
sudo systemctl start netd

To configure netd to start when the box boots:
sudo systemctl enable netd"

If I do start the netd and configure it to run at start it will sometimes not complain but it doesn't make any difference in showing or saving configuration.

Sometimes it does complain like below:

"cumulus@sw01-nni-1:~$ sudo systemctl start netd
Job for netd.service failed. See 'systemctl status netd.service' and 'journalctl -xn' for details."

"cumulus@sw01-nni-1:~$ systemctl status netd.service
â netd.service - Network Command Line Utility Daemon
Loaded: loaded (/lib/systemd/system/netd.service; enabled)
Active: failed (Result: start-limit) since Wed 2017-05-03 09:03:16 UTC; 3min 9s ago
Process: 841 ExecStart=/usr/bin/python /usr/sbin/netd -d (code=exited, status=1/FAILURE)
Main PID: 841 (code=exited, status=1/FAILURE)"

"cumulus@sw01-nni-1:~$ journalctl -xn
No journal files were found."

I can start netd again and it comes alive but then fails when trying to run net commands.

"cumulus@sw01-nni-1:~$ sudo systemctl start netd
cumulus@sw01-nni-1:~$ systemctl status netd.service
â netd.service - Network Command Line Utility Daemon
Loaded: loaded (/lib/systemd/system/netd.service; enabled)
Active: active (running) since Wed 2017-05-03 09:14:41 UTC; 2s ago
Main PID: 939 (python)
CGroup: /system.slice/netd.service
ââ939 /usr/bin/python /usr/sbin/netd -d"

I can't find any bug pertaining to cumulus VX so I suspect it maybe due to the EVE-NG cumulus support and perhaps some update needed to the way the 3.2.1 image is supported in EVE?

Cheers

[ramindia]

Hi

I have used 3.X before i did not see any issue, but your version is different.

when i get chance i will try and let you know. ( not so soon due to other projects).

R!

[Sipart]

Thanks R! I will try an older 3.x version and see how it goes.

Cheers

[Sipart]

Cumulus recently released 3.3.0 of the VX, so I tried to upgrade the switch but it failed to so I backed up the config and added the new image and it works great.

[ramindia]

Cumulus recently released 3.3.0 of the VX, so I tried to upgrade the switch but it failed to so I backed up the config and added the new image and it works great.

good to know it worked.

R!

[powen]

Out of curiosity, how much RAM did you give the 3.2.1 instance?

I am unsure if this is related, however, the EVE-NG default Cumulus VX template suggests 256MB. The minimum should be 512MB. NCLU wouldn't work for me at 256MB, yet Python was throwing some fairly suggestive error messages hinting at the problem. An upgrade to 512MB per instance saw all problems go away.

[Sipart]

Out of curiosity, how much RAM did you give the 3.2.1 instance?

I am unsure if this is related, however, the EVE-NG default Cumulus VX template suggests 256MB. The minimum should be 512MB. NCLU wouldn't work for me at 256MB, yet Python was throwing some fairly suggestive error messages hinting at the problem. An upgrade to 512MB per instance saw all problems go away.

To be honest I can't remember as I destroyed the orignal node but on the new ones I used 512MB or 1G.

So you could of been right

Cheers