Hi, I am going to change some IPs of servers in the cage, and do some cleanup.
We are using now 4 /29 of the /24 we have in the cage (3 contiguous, and 1 separated), and after some quick math, I think we could go down to half of it. For that, I would need to: - change hypervisors IP to not use a isolated /29. This is like this for historical reason, but we can move them. This will not impact anything, except if I get it wrong during the move. This should free a /29 (8 IPs) - remove old services. As people know, we are in the process of moving out of gerrit for the last users. Once this part is done, we will be able to - remove gerrit (it required a public IP for git push) - remove the stage gerrit - remove the postgresl server (needed for gerrit, so in the same vlan) so 3 IPs. Get rid of some test server (chrono.rht) Move more services in the lan: - softserve - fstat - jenkins - jenkins stage - bugziller I think softserve, bugziller and fstat are already there, but we kept the old server to rollback in case of problems. I guess that it is time to stop them (and verify the migration was correct). Jenkins is not going to be moved easily so I keep that for the future. Jenkins stage is currently down so it can be reinstalled elsewhere without trouble. This would free 4/5 IPs. Finally, the trickiest part is that our firewalls and reverse proxy each use 3 IP. 1 floating, and 2 for each side of the pair. There is no reason to have a public IP for each, they just need the floating IP. This would free 4 IPs, but that's pretty high risk, so this will happen likely last. All of this to say that if you see any weird network issue for anything the cage, please tell us. For now, I just touched to 1 hypervisor, the least risky one. -- Michael Scherer / He/Il/Er/Él Sysadmin, Community Infrastructure
signature.asc
Description: This is a digitally signed message part
_______________________________________________ Gluster-infra mailing list Gluster-infra@gluster.org https://lists.gluster.org/mailman/listinfo/gluster-infra