Hi,
if you value your data you shouldn't set it to k (in your case 6). The
docs [1] are pretty clear about that:
min_size
Sets the minimum number of replicas required for I/O. See Set the
Number of Object Replicas for further details. In the case of
Erasure Coded pools this should be set to a value greater than ‘k’
since if we allow IO at the value ‘k’ there is no redundancy and
data will be lost in the event of a permanent OSD failure.
Regards,
Eugen
[1] https://docs.ceph.com/en/latest/rados/operations/pools/#set-pool-values
Zitat von Christopher Durham <caduceus42@xxxxxxx>:
Hello,
I have an ec pool set as 6+2. I have noticed that when rebooting
servers during system upgrades, I get pgs set to inactive while the
osds are down. I then discovered that my min_size for the pool is
set to 7, which makes sense that if I reboot two servers that host a
pg with OSDs on both servers, then only 6 of the OSDs are available
during the reboot cycle, and with a min_size of 7, they go inactive.
Is is ok for me to set min_size on the pool to 6, to avoid the
inactive problem? I know I could do my reboots sequentially to
eliminate multiple server downtime, but wanted to be sure the
min_size 6 is ok. I know this may increase other risks, but wanted
to know if this min_size change is an option, albeit more risky.
Thanks.
-Chris
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
