[ Blog ] Shutdown #Proxmox #cluster with #Ceph storage
To #shutdown Proxmox cluster and prevent data loss or corruption, especially when Ceph storage is in use, you must follow a specific procedure.
When working with a high-availability infrastructure, performing maintenance tasks can be a delicate process. Shutting down a cluster, even for a planned event like a power outage or hardware http://rviv.ly/GCyTZM
@korkenzieher OK, that was a really helpful RTFM. I now have a virtual IP up and running and can connect to the S3 endpoint via that vIP.
Sadly, the TLS settings seem to be a little less capably in the tentacle release. If I understand the documentation (https://docs.ceph.com/en/tentacle/cephadm/services/rgw/#high-availability-service-for-rgw) correctly, I can handover a TLS certificate and key.
In the `latest` documentation on can specify this should be a self-signed certificate provided by cephadm.
So until I get the Let's Encrypt certificate, I have HTTP traffic served on port 443, as just specifying `ssl: true` seems to do nothing in Tentacle...
@korkenzieher Sometimes I do not understand the Ceph documentation structure. Why are there two codeblocks in that section, being almost identical, but without an explanation what the difference is?
I found that there are two instances of the RGW service running, which are listening on port 80. I managed to put them behind a simple haproxy (that was already in my homelab) and it seems to work.
Having a virtual IP on all three nodes that hosts the loadbalancer, similar to how kolla-ansible is doing it for the OpenStack things, would be really nice. But I do not want to build and maintain that myself and was hoping for something out of the box...
OK, I am officially confused by the ceph and cephadm documentation (mostly the fact that some things are in the cephadm one and others are in the "general" documentation)
I managed to get a three-node cluster with some old hardware running (as a demo, not productive). MGRs are running, MONs are running, OSDs are running. RGWs are running. The ceph dashboard is running.
Can anyone point me in the right direction / RTFM on how to supply proper certificates to a ceph cluster installed with cephadm? Mostly for the ceph dashboard and the S3 endpoints.
On a related note, do I need to setup haproxy / keepalived / ... myself to make the cluster "highly available" (in my case just surviving the reboot of a node, just to show the setup is "working")? Or can cephadm or similar do that for me, too?
[ Blog ] Proxmox upgrade #Ceph #Reef to #Squid
If you are running Ceph Reef in your #Proxmox infrastructure and plan to upgrade to Proxmox 9, you must first upgrade Ceph Reef to Squid to meet the prerequisites.
As a best practice, before proceeding with the upgrade make sure to have a working backup of your VMs and Containers.
Prerequisites
To upgrade Ceph Reef to http://rviv.ly/tmWeXQ #aggiornamento
[ Blog ] Shutdown #Proxmox #cluster with #Ceph storage
To #shutdown Proxmox cluster and prevent data loss or corruption, especially when Ceph storage is in use, you must follow a specific procedure.
When working with a high-availability infrastructure, performing maintenance tasks can be a delicate process. Shutting down a cluster, even for a planned event like a power outage or hardware http://rviv.ly/GCyTZM
[ Blog ] Remove #Ceph #datastore from #Proxmox
To remove Ceph datastore from Proxmox several steps are involved, especially if you want to completely clear Ceph from your system.
Before proceeding, keep in mind that this process will permanently delete all data stored on the Ceph datastore. Make sure to have a working backup of any critical data before proceeding.
Migrate VM disks off http://rviv.ly/9WJn4b
[ Blog ] Proxmox #node replacement in #Ceph cluster
If a core Proxmox server fails taking its Ceph OSDs with it, the Proxmox node replacement doesn't have to be a nightmare.
To fix this issue you must cleanly decommission the failed server and correctly perform the Proxmox node replacement to ensure your Ceph data remains resilient.
The Ceph cluster should have the status http://rviv.ly/hDkYPo #replace
Paolo Valsecchi
Johannes Kastl
Yehor ๐บ๐ฆ