Discussion thread for official Zabbix SMART Disk monitoring

[rob4ikomg]

Hello all

A have host "Windows 10" :
--- zabbix agent (Active). Version: "zabbix_agentd Win64 (service) (Zabbix) 6.2.3". Location "C:\monitoring\zabbix\zabbix_agent.conf".
--- Smart tool "smartmontools". Version: "smartctl 7.3 2022-02-28 r5338 [x86_64-w64-mingw32-w10-21H2] (sf-7.3-1)". Location: "C:\Program Files\smartmontools\bin\smartctl.exe".

zabbix agent config:
LogFile=c:\monitoring\zabbix\zabbix_agentd.log
DebugLevel=3
LogFileSize=1

StartAgents=0

ServerActive=178.151.*.*:*
Hostname=home
RefreshActiveChecks=61

Timeout=30
​

I added a host to the zabbix server, but I get an error:

zabbix_agent.log

  7364:20221016:074123.806 Starting Zabbix Agent [home]. Zabbix 6.2.3 (revision 98ee88fc19d).
  7364:20221016:074123.807 **** Enabled features ****
  7364:20221016:074123.807 IPv6 support:          YES
  7364:20221016:074123.807 TLS support:           YES
  7364:20221016:074123.808 **************************
  7364:20221016:074123.808 using configuration file: C:\monitoring\zabbix\zabbix_agent.conf
  7364:20221016:074124.251 agent #0 started [main process]
 14892:20221016:074124.252 agent #1 started [collector]
 16404:20221016:074124.252 agent #2 started [active checks #1]
 16404:20221016:074138.565 active check "smart.disk.discovery" is not supported: Unsupported item key.
 16404:20221016:074148.857 active check "smart.disk.discovery" is not supported: Unsupported item key.
 16404:20221016:074158.991 active check "smart.disk.discovery" is not supported: Unsupported item key.
 16404:20221016:074208.097 active check "smart.disk.discovery" is not supported: Unsupported item key.​

what am I doing wrong?

[Ch77]

I added SMART Template at the second Debian server and I received this

Cannot create item: item with the same key "smart.disk.get[{#PATH},"{#RAIDTYPE}"]" already exists.
Cannot create item: item with the same key "smart.disk.get[{#PATH},"{#RAIDTYPE}"]" already exists.
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".

However there is no items with key such "smart.*"

Problem was in zabbix agent version. I upgrade from 5.4 to 6.4 and problem was gone

[Ch77]

I added SMART Template at the second Debian server and I received this

Cannot create item: item with the same key "smart.disk.get[{#PATH},"{#RAIDTYPE}"]" already exists.
Cannot create item: item with the same key "smart.disk.get[{#PATH},"{#RAIDTYPE}"]" already exists.
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".
Cannot accurately apply filter: no value received for macro "{#ATTRIBUTES}".

However there is no items with key such "smart.*"

[doogie]

Can somebody explain to me what it means? I have a trigger error "SMART [sda]: Some command to the disk failed" and Exit status - 4 - "Bit 4: We found prefail Attributes <= threshold."

I'm getting this error as well, but only on Debian servers for some reason. All the CentOS boxes seem to be working fine. They're all the same - Dell servers with Megaraid and all seems fine from smartctl side.

[Ch77]

Can somebody explain to me what it means? I have a trigger error "SMART [sda]: Some command to the disk failed" and Exit status - 4 - "Bit 4: We found prefail Attributes <= threshold."

[mhk]

I have a problem that seems to be purely related to Proxmox VE servers (version 6.4 and 7.x as reported by a user in a previous post).

same here.
the thing I found in regards to Proxmox - that I can't find a way to undo - is this:

from the agent log:

2022/08/23 15:30:52.485456 executing direct exporter task for key 'smart.disk.discovery'
2022/08/23 15:30:52.485476 [Smart] executing smartctl command: sudo -n smartctl --scan -j
2022/08/23 15:30:52.485900 [Smart] command sudo -n smartctl --scan -j smartctl raw response: sh: 1: sudo: not found
2022/08/23 15:30:52.485926 failed to execute direct exporter task for key 'smart.disk.discovery' error: 'Cannot fetch data: Failed to scan for devices: Cannot unmarshal JSON: invalid character 's' looking for beginning of value..'

Notice the "sudo -n smartctl --scan [...]" and then the "sh: 1: sudo: not found"
which from my assumption is what causes it, since Proxmox runs on Debian as root (no sudo pre-installed) but the plugin always want's to call the discovery with sudo and hence fails...

any idea (apart from installing sudo, which I really don't want tbh) how to tell the thing to call it's function without sudo?

at least that's what I found out so far, trying to solve this

ZBX_NOTSUPPORTED: Cannot fetch data: Failed to scan for devices: Cannot unmarshal JSON: invalid character 's' looking for beginning of value..

[bcnx]

I have a problem that seems to be purely related to Proxmox VE servers (version 6.4 and 7.x as reported by a user in a previous post). The problem becomes apparent wit this command:

/usr/local/sbin/zabbix_agent2 -c /usr/local/etc/zabbix_agent2.conf -t smart.disk.discovery
smart.disk.discovery [m|ZBX_NOTSUPPORTED] [Failed to execute smartctl: Command execution failed: exit status 127.]

The Zabbix Agent 2 is running as root, so no permission problems. Using smartmontools 7.1.

I did a ton of troubleshooting, but to no avail. Any input is appreciated.

Cheers

BC

[bcnx]

You can use a parameter in the config file for the Zabbix Agent 2. You can point to the actual position of your smartctl binary. Less messy than a symlink.

[bcnx]

I have the exact same problem on Proxmox VE 6.4. It works fine on Ubuntu. No idea why, I'm using an alternative way of NMVe monitoring until I find a solution.

[bcnx]

I think I know the problem: these commands output a very different output. It seems NVMe drives are less suitable to be monitored by SMART:

# smartctl -a /dev/nvme0 -j
# smartctl -a /dev/sda -j

Not sure how to have NVMes output the same info. Maybe I need to monitor those by using nvme-cli (https://github.com/narbehaj/zabbix-nvme)

Cheers,

BC