Thanks! Very good links! :)
I need to subtract from the usable capacity max usable/server count to handle 1 server failure. Anything else I need to subtract?
That buffer for server failure recovery is a good idea and often missed. These days though Ceph is pretty good at detecting that an entire node is down, and with a judicious mon_osd_down_out_subtree_limit one can usually avoid a thundering herd if the node can be brought back up within a short period of time. That said, it’s still good to have that margin in case it takes a long time. The backfillfull/full ratios can be squeezed higher in a pinch, but that can be risky.
https://docs.clyso.com/tools/erasure-coding-calculator/
Am Sa., 16. Nov. 2024 um 10:04 Uhr schrieb Marc Schoechlin <ms@256bit.org>:
Hi Anthony,
this is a nice one! The original mail had a broken link :-) https://www.osnexus.com/ceph-designer
Curious. Apple Mail has been weird on me lately.
Perhaps this would also be a good project to build an open platform that helps to do a dteailed planning of Ceph clusters in general (server and switch hardware, different types of pools, service distribution, placement groups, ....). Especially if it also had the option of using or creating hardware profiles (that the user creates himself or that interested hardware manufacturers can create) from different server and switch manufacturers and comparing them in terms of costs, energy consumption, throughput and usable memory.
Regards Marc
Am 15.11.24 um 16:55 schrieb Anthony D'Atri:
https://www.osnexus.com/ceph-designer
On Nov 15, 2024, at 10:51 AM, Marc<Marc@f1-outsourcing.eu> wrote:
I was wondering if there is some online tool that can help you with calculating usable storage from nodes, drives per node, replication/erasure used etc.
_______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io