Ceph s3 and swift rgw

[Leo David]

Hello everyone,
I've searched on the forum, but could't find something related to rgw features of ceph ( pve4 & jewel ). Maybe I haven't searched enough, but is that posible o be configured under proxmox environment ?
I would like to benefit of these features along all the good things that pve brings...
Thank you !

 

[fabian]

no, PVE does not use or setup RGW in any way.

 

[Leo David]

OK. But manually configuring and adding a rgw an generally editing ceph.conf would it be ok ?

 

[fabian]

OK. But manually configuring and adding a rgw an generally editing ceph.conf would it be ok ?

If you know what you are doing, you are free to edit the ceph config. Thoroughly testing any such changes is of course advisable, but in general PVE will not change much about the configuration after the initial setup (besides adding monitor entries when creating monitors for example).

 

[Leo David]

If you know what you are doing, you are free to edit the ceph config. Thoroughly testing any such changes is of course advisable, but in general PVE will not change much about the configuration after the initial setup (besides adding monitor entries when creating monitors for example).

Thank you.

If you know what you are doing, you are free to edit the ceph config. Thoroughly testing any such changes is of course advisable, but in general PVE will not change much about the configuration after the initial setup (besides adding monitor entries when creating monitors for example).

Thank you. So in generally, after a succesfull setup of ceph, if I install ceph-deploy on one of the machines and I'll further manage the cluster as from a ceph-admin node with this utility it won't create issues.

 

[fabian]

Thank you.

Thank you. So in generally, after a succesfull setup of ceph, if I install ceph-deploy on one of the machines and I'll further manage the cluster as from a ceph-admin node with this utility it won't create issues.

ceph-deploy does a lot of "automagical" configuration, so this is not possible to say in general. I'd advise to stick to manual configuration steps to ensure you fully understand what is happening (and what consequences are implied). that said, PVE mainly cares about being able to access the configured storage pools, so the most important thing is to have the right keyrings in the right places, and that ceph.conf is consistent cluster-wide.