Preview, Feedback wanted - Cluster Dashboard

[Ashley]

Updated CEPH to Jewel which splits Read & Write IOPS into separate values, in the dashboard it splits this into two outputs correctly, however write always shows 0, read shows the correct read value reported in my other tools.

I know currently Jewel is not supported in Proxmox but thought was worth noting as I know is around the corner.

 

[dcsapak]

Updated CEPH to Jewel which splits Read & Write IOPS into separate values, in the dashboard it splits this into two outputs correctly, however write always shows 0, read shows the correct read value reported in my other tools.

this should work...

can you send me the output of "ceph status -f json-pretty" (per pm if you dont want to post it here)

 

[Ashley]

this should work...

can you send me the output of "ceph status -f json-pretty" (per pm if you dont want to post it here)

Sent via PM

 

[dcsapak]

Sent via PM

thanks, fix is already on the list, it was just a typo

 

[frantek]

How is the dashboard storage calculated by the API call it uses? I see a relatively big difference between dashboard und Ceph. Dashboard reports at my cluster 2.62 TiB of 11.29 TIB used and Ceph 23.59 Gib of 8.18 TiB.

 

[dcsapak]

How is the dashboard storage calculated by the API call it uses? I see a relatively big difference between dashboard und Ceph. Dashboard reports at my cluster 2.62 TiB of 11.29 TIB used and Ceph 23.59 Gib of 8.18 TiB.

all combined (all except local are only counted once)

 

[frantek]

Suggestion: Add storage details to the dashboard like the node lists which lists all types of storage (for me: ceph, nfs backup, 3 x local) its size, usage and a sum over all items in the list.

 

[dcsapak]

my plan is to be able to select storages which are counted towards the total, but your suggestion makes also sense, but i have to see what works better