Script fails if spare disk is in use

[lonnie44]

Hi guys,

My zpool looks as follows: (replaced IDs)

NAME                           STATE     READ WRITE CKSUM
tank                           DEGRADED     0     0     0
raidz2-0                       DEGRADED     0     0     0
  sda                          ONLINE       0     0     0
  sdb                          ONLINE       0     0     0
  sdc                          ONLINE       0     0     0
  spare-3                      DEGRADED     0     0     0
    sdd                        FAULTED     34     0     0  too many errors
    sdf                        ONLINE       0     0     0
  sde                          ONLINE       0     0     0
spares
  sdf                          INUSE     currently in use

But Zabbix did not trigger because the script failed with this:

# /etc/zabbix/scripts/zfs.py
Traceback (most recent call last):
  File "/etc/zabbix/scripts/zfs.py", line 155, in <module>
    scrub, vdev_errors = pool_status()
  File "/etc/zabbix/scripts/zfs.py", line 78, in pool_status
    vdev_errors = {x[0]: {
  File "/etc/zabbix/scripts/zfs.py", line 79, in <dictcomp>
    'read': int(x[2]),
ValueError: invalid literal for int() with base 10: 'currently'

https://github.com/blind-oracle/zabbix-zfs/blob/18d396c9c0148881c70f506f37939ad5f8827396/zfs.py#L78-L82

'r' contains the following and will fail to convert to int(): [... ['/dev/sde1', 'ONLINE', '0', '0', '0'] ['spares'], ['/dev/sdf1', 'INUSE', 'currently', 'in', 'use'] ...]

Would be great if someone could fix that.

Best regards

[blind-oracle]

Yeah I've never tested it with spares as I don't use them :) Will try to adapt the script

[blind-oracle]

I think I've fixed it by skipping whatever cannot be converted, please check

[lonnie44]

Still fails but different:

Traceback (most recent call last):
File "/etc/zabbix/scripts/zfs.py", line 170, in <module>
'vdevs': vdev_list(vdev_errors),
File "/etc/zabbix/scripts/zfs.py", line 106, in vdev_list
return {x[0]: {
File "/etc/zabbix/scripts/zfs.py", line 114, in <dictcomp>
'errors': errors[x[0]],
KeyError: '/dev/sdd1'

I think this is the issue, because of the text at the end 'too many errors'. This should have been a issue before as well. https://github.com/blind-oracle/zabbix-zfs/blob/0779045794f67a41c8b7d3aa17c78c5e9c04eb29/zfs.py#L83-L85

[lonnie44]

Fixed it, it is working now: I replaced

if len(x) != 5 or not x[0].startswith('/'):

with

if len(x) < 5 or not x[0].startswith('/'):

Since all convert errors are skipped I guess it is fine.

Thanks a lot!

[blind-oracle]

Ok, good. I'll try to convert the script at some point to use zfs program with lua and JSON, it will probably make the implementation stable and not dependant on the zpool/zfs output