Ceph Monitor address already in use (500) error

[yyyy]

Hello,

for some reason it keeps showing the error that the monitor address is already in use (500) in the ceph Monitor configurator, the list shows there is no such address being used.

 

[sb-jw]

What exactly are you trying to do that you may have done before?

 

[yyyy]

What exactly are you trying to do that you may have done before?

I removed the old Node (.189.0) and it disappeared from the list, it was a ghost monitor which kept showing offline despite the actual node being accessible therefore I wanted to remove it an re-add it, followed this guide https://forum.proxmox.com/threads/ghost-monitor-in-ceph-cluster.58683/

but although it disappeared from the ceph monitor list, it appears that I can no longer add it again

 

[sb-jw]

A .0 address is rather unusual, are you sure that's correct? Or do you actually use networks larger than /24?

 

[yyyy]

A .0 address is rather unusual, are you sure that's correct? Or do you actually use networks larger than /24?

Yes I am on /16 subnet for larger address space

 

[sb-jw]

I would strongly advise against this, especially for CEPH. You should limit the broadcast domain to a minimum and a /24 for CEPH is usually more than enough. Otherwise, you have a lot of unnecessary traffic in the network which take up resources for storage. In addition, disruptions, loss of traffic or complete standstill can occur at some point.

How exactly did you remove the mon? Have you made sure nothing was left behind?

 

[yyyy]

I would strongly advise against this, especially for CEPH. You should limit the broadcast domain to a minimum and a /24 for CEPH is usually more than enough. Otherwise, you have a lot of unnecessary traffic in the network which take up resources for storage. In addition, disruptions, loss of traffic or complete standstill can occur at some point.

How exactly did you remove the mon? Have you made sure nothing was left behind?

Sure but I don't think this is the issue, it is an IP so if it cannot handle this then there is something fundmentally wrong with Proxmox's implementation, somewhere something is broken. I removed the mon by following the guide as I said : https://forum.proxmox.com/threads/ghost-monitor-in-ceph-cluster.58683/

 

[sb-jw]

But I didn't say that the /16 network was a problem for Proxmox. But it is a fundamental architectural flaw in the network design, which will lead you into serious problems sooner or later. In the /16 everyone wants information about someone else, and over time your background noise will take up a significant portion of your bandwidth.

But well, it's up to you.

I removed the mon by following the guide as I said : https://forum.proxmox.com/threads/ghost-monitor-in-ceph-cluster.58683/

Well, that's how it only helps to a limited extent. On the one hand, I don't want to read through old threads to put myself in your problem and, on the other hand, your general conditions, error messages or commands were different, you have to share this information and not refer to another thread.
No Front, but we're all here to help you voluntarily, but in return we can expect to get the information from you and not have to put it all together and ask our crystal ball about it.