the default min_size for an EC pool is always k + 1, except when you choose m = 1, then it will be k (I won’t go into details if m = 1 is a good choice or not)
I will. It’s a bad idea for data you can’t easily reconstruct or write off. Indeed if you need continued availability if 3 hosts out of 9 are down, I would just do replica3. You could do a custom or MSR CRUSH rule but that’s a rabbit hole. With Umbrella you should be able to transparently migrate pools to different protection schemes … one it gets a lot of shakeout.
. You can temporarily reduce min_size to k, in a disaster recovery scenario it can help. But k + 1 is a safety switch, so if you reduce it, ensure to set it back after recovery. If you plan to take multiple hosts offline on a regular basis, I’d recommend to consider a different EC profile. Note that you can’t change the EC profile of a pool, you would need to create a new pool with the desired profile and copy the data over to the new pool.
Regards Eugen
Zitat von "GLE, Vivien via ceph-users" <ceph-users@ceph.io>:
Hi,
We have a 9 hosts (3 per pod) cluster with 1 pool in EC 5+3 (host domain failure), I expected that shutting down 1 pod wouldn't affect our pg but I just discovered that the EC pool is in min_size 6 while trying it.
Is min_size 6 important in this case ? Could I safely lower it to 5 or using another EC k+m would suit better ?
Thanks !
Vivien _______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io
_______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io