Question

poed is not responding


Userlevel 2
trying to check poe status of an attached AP, but running poectl -s returns errors stating that the PoE Daemon isn't running. Trying to the start the service doesn't work.

cumulus@REDACTED:mgmt-vrf:~$ sudo systemctl status cumulus-poe.service
● cumulus-poe.service - Cumulus Linux POE Daemon
Loaded: loaded (/lib/systemd/system/cumulus-poe.service; enabled)
Active: active (running) since Mon 2019-01-14 15:12:13 UTC; 9s ago
Main PID: 5918 (poed)
CGroup: /system.slice/cumulus-poe.service
└─5918 /usr/bin/python /usr/sbin/poed

Jan 14 15:12:13 DSW41-BKLNNY systemd[1]: Started Cumulus Linux POE Daemon.
Jan 14 15:12:13 DSW41-BKLNNY poed[5918]: Starting POE Daemon
cumulus@REDACTED:mgmt-vrf:~$ sudo poectl -s
Error: unable to get POE status. Is poed daemon running?
cumulus@REDACTED:mgmt-vrf:~$ sudo systemctl start poed
Failed to start poed.service: Unit poed.service failed to load: No such file or directory.
cumulus@REDACTED:mgmt-vrf:~$

Over the weekend the switch became un-responsive, though it was still had power. Field tech performed a power-cycle on the switch, which brought it back up, and second PoE Wireless AP on the switch is pulling power and is online, but poectl isn't responding.
Any insight?

4 replies

Userlevel 2
Hi Troy,

Which version of Cumulus Linux are you running?
Was this system recently upgraded from an older version of Cumulus Linux?
Can we capture the output from the following two commands?
code:
cat /etc/lsb-release
cat /etc/image-release


In the past, the poed package was replaced with the cumulus-poe package. There was a condition during upgrades between versions while using APT where both packages could be present on the switch, and enabled. To verify if both packages are installed on your switch, please run the command:
code:
dpkg -l | grep poe


Only one poe package should be present on the switch:
code:
$ dpkg -l | grep poe
ii cumulus-poe 3.0.1-cl3u11 all Cumulus Linux Power Over Ethernet (POE) tools and service.


If both packages are present, it's likely a conflict of both being installed that is resulting in the behavior you are seeing.
Userlevel 2
cat /etc/lsb-release:
DISTRIB_ID="Cumulus Linux"
DISTRIB_RELEASE=3.6.2
DISTRIB_DESCRIPTION="Cumulus Linux 3.6.2"

cat /etc/image-release:
IMAGE_ID="Cumulus Linux"
IMAGE_RELEASE=3.6.1
IMAGE_DESCRIPTION="Cumulus Linux 3.6.1"
IMAGE_SWITCH_ARCHITECTURE=bcm-armel
IMAGE_BUILD_ID=a89d37cz7e7fd32
IMAGE_BUILD_DATE=2018-05-25T22:46:21+0000
IMAGE_BUILD_USER=clbuilder

and running just the cumulus-poe package 3.0.1-c13u11

running systemctl status cumulus-poe returns:
● cumulus-poe.service - Cumulus Linux POE Daemon
Loaded: loaded (/lib/systemd/system/cumulus-poe.service; enabled)
Active: active (running) since Tue 2019-01-15 02:33:14 UTC; 11h ago
Main PID: 1446 (poed)
CGroup: /system.slice/cumulus-poe.service
└─1446 /usr/bin/python /usr/sbin/poed

Jan 15 02:33:14 REDACTED systemd[1]: Started Cumulus Linux POE Daemon.
Jan 15 02:33:16 REDACTED poed[1446]: Starting POE Daemon
Jan 15 02:33:26 REDACTED poed[1446]: ERROR : No response from POE hardware.

So if the PoE module has faulted, how would a PoE device on swp1 be receiving power?
This is on an Edge-Core Accton AS4610-54P switch.
Userlevel 2
Hi Troy,

I can't provide any speculation as to why the PoE module would still be able to provide power although the hardware itself seems to have failed. Does this behavior persist through a power-cycle of the switch?

Remove power from the switch
Wait 30 seconds
Restore power to the switch



If the behavior persists, would recommend next steps of pursuing an RMA with the vendor where the hardware was purchased.
Userlevel 2
Did both hard power cycle and software reboot, still no PoE beyond swp1. Looks like its time to punt this over the manufacturer.

Reply