Originally from the User Slack
@Cong_Guo: Hi experts, the list of replace_address_first_boot in my ScyllaCluster CR seems messed up in my deployment, do you know how to modify the list below? like how could I delete scylla-ads-b-20 below ?
status:
observedGeneration: 13
racks:
b:
conditions:
- status: "True"
type: RackUpgrading
- status: "True"
type: MemberReplacing
members: 35
readyMembers: 30
replace_address_first_boot:
scylla-ads-b-20: 10.0.9.251
stale: false
updatedMembers: 31
@Maciej_Zimnoch: it’s internal field, you shouldn’t mess with it. Replace was initiated on node with 20 suffix, check if it’s proceeding and if it’s correct IP address.
I guess you’re using outdated operator version which is no longer supported, as we no longer use this field.
@Cong_Guo: Yes, it’s an old version operator which we plan to upgrade. The ip to be replaced is incorrect for it is a live node in another DC.
@Maciej_Zimnoch: If it’s incorrect and replace is not ongoing, to clean it up you can patch CRD subresource status using kubectl patch with --subresource=status parameter. I don’t have an example so refer to documentation and try only touch the replace_address_first_boot, don’t change other fields.
--dryrun=server might be useful as well to test it out.
@Cong_Guo: Got it, thanks a lot for your help!
> If it’s incorrect and replace is not ongoing, to clean it up you can patch CRD subresource status using kubectl patch with --subresource=status parameter.
It works really well, thank you very much!