Ceph_df broken for ceph 15.2 (octopus)

Troubleshooting
#1

Hi All,

i found the check ceph_df is broken for Ceph 15.2 Octopus.

Since format of ceph df detail has changed, the output is no loger recognized by the check ceph_df.

The actual output for the sections is:

— RAW STORAGE —
CLASS SIZE AVAIL USED RAW USED %RAW USED
xxxx 15 TiB 11 TiB 4.2 TiB 4.3 TiB 28.32
yyyy 20 TiB 14 TiB 5.3 TiB 5.3 TiB 27.29
TOTAL 35 TiB 25 TiB 9.6 TiB 9.6 TiB 27.74

— POOLS —
POOL ID STORED (DATA) (OMAP) OBJECTS USED (DATA) (OMAP) %USED MAX AVAIL QUOTA OBJECTS QUOTA BYTES DIRTY USED COMPR UNDER COMPR
cephfs_data 2 97 GiB 97 GiB 0 B 24.91k 97 GiB 97 GiB 0 B 1.31 3.6 TiB N/A N/A 24.91k 0 B 0 B
cephfs_metadata 3 5.5 MiB 5.5 MiB 4.7 KiB 24 5.5 MiB 5.5 MiB 4.7 KiB 0 3.6 TiB N/A N/A 24 0 B 0 B
xxxx 9 2.7 TiB 2.7 TiB 1.6 KiB 696.59k 2.7 TiB 2.7 TiB 1.6 KiB 19.16 5.6 TiB N/A N/A 696.59k 0 B 0 B
yyyy 10 2.0 TiB 2.0 TiB 4.9 KiB 539.08k 2.0 TiB 2.0 TiB 4.9 KiB 22.12 3.6 TiB N/A N/A 539.08k 0 B 0 B
device_health_metrics 12 0 B 0 B 0 B 6 0 B 0 B 0 B 0 5.5 TiB N/A N/A 6 0 B 0 B

so the sections and headers are change since version 15 of ceph.

Need help to change, update and test the check for Open Monitoring Distribution Version 1.6.0p19.cre

Regards Lukas

1 Like
#2

Hi Lukas,

I’m experiencing the same issue and could need help also.
Here are some additional informations:
We have 3 Ceph-Clusters in different Versions.
The Check_MK-Server version is 1.6.0p20.
The Agents are the same version.

Cluster 1:
Ceph-Version 14.2.11 has no problems with ceph_df.

Cluster 2:
Ceph-Version 14.2.16 the check can’t find ‘Ceph Pool POOLNAME’ because ‘UNKN - Item not found in agent output’.

Cluster 3:
Ceph-Version 15.2.8 the check can’t find ‘Ceph Pool SUMMARY’ but also ‘Ceph MGRs’ because ‘UNKN - Item not found in agent output.’

In all versions the agent output ‘<<<ceph_df>>>’ differs:
##############################
Version 14.2.11:
POOLS:
POOL ID STORED OBJECTS USED %USED MAX AVAIL QUOTA OBJECTS QUOTA BYTES DIRTY USED COMPR UNDER COMPR
##############################
Version 14.2.16:
POOLS:
POOL ID PGS STORED OBJECTS USED %USED MAX AVAIL QUOTA OBJECTS QUOTA BYTES DIRTY USED COMPR UNDER COMPR
##############################
Version 15.2.8:
— POOLS —
POOL ID PGS STORED (DATA) (OMAP) OBJECTS USED (DATA) (OMAP) %USED MAX AVAIL QUOTA OBJECTS QUOTA BYTES DIRTY USED COMPR UNDER COMPR
##############################

For the ‘Ceph MGRs’ I coulnd’t find the correct position in the agent output. But I saw that ‘<<<ceph_status>>>’ has less output than the lower versions.

Please tell me if you need more Informations.

Regards,
Marc-Oliver

#3

Hey @mensinck,

according to tribe29 the bug is fixxed in Check_MK Version 2.0.

Kind Regards,
Marc-Oliver

#4

Hi, sorry to pull this up again after six months.

We also have the problem, that check_mk got problems with the Ceph MGRs service. We did the update to the lastet check_mk 2.0.0p12 CRE version.

Is there any way to fix that?

#5

Hi @KervyN on my site we also still have problems with Ceph MGRs service.
The other ceph services are working fine…

Currently we have no solution for this behavior.
Using check_mk 2.0.0p6 CRE version.

#6

@mwernet oh joy.

Is there a bug somewhere already open for this?

#7

Hi All,

We have upgraded to ceph 16.2.6 in the meantime.
This resolved the ,Ceph Pool Summery" and ,Ceph Pool POOLNAME" problems with ceph_df

image
image1315×59 7.67 KB

Ceph MGR’s is still ,UNKN Item" but this is ,ceph_status_mgrs" as I see. This is not covered by mk_ceph.

Regards Lukas