We're about to use nsq in openshift, and found these manifests:
I have posted a comment to the author, but he didn't answer:
"Hey, I really don't get how this could work with several replicas. Could you explain how you managed to notify all the (3 replicas of) nslookupd when nsqd is starting?
Also pod IPs are considered ephemeral, what IP do the 5 replicas of nsqd register? If it's the ip of the service, what's the point of having nslookupd?
Finally, the 5 nsqd replicas are sharing a single PV, this will lead to io error as soon as 2 servers will write at the same place, how do you handle that?
Thanks for your insights."
Am I understanding something wrong?
I don't think we can have one claim PER replica, right?