Ancestors

Toot

Written by Rachel on 2025-01-22 at 18:22

Cluster rebuild project:

Ok due to hiding in the woods I haven't made much progress here

Current status:

• Will soon test power usage of stripped poweredge Proxmox servers, if successful my node count issues are mostly moot
• Not sure if a single ceph node on 1gb with 3 others on 10gbe is an issue?
• Coming around to the idea of using either terraform or cluster-api to build the new cluster will evaluate more

=> More informations about this toot | More toots from rachel@transitory.social

Descendants

Written by Michael on 2025-01-22 at 18:58

@rachel I think the 1 gb node will at least be an issue for writes, depending on how loaded the connection to that node is, because a write in replicated pools is only acknowledged to the client once all copies of the data have been written.

I'm also very curious to see what your experience working with cluster-api will look like. OnceI'm done with the migration, I want to have another look at how I manage the underlying OS. I'm currently thinking about Canonical's Metal-as-a-service thing.

=> More informations about this toot | More toots from mmeier@social.mei-home.net