We have a new load balancer layer cloudlb that can be used to expose services to the WMCS realm from hardware in the production realm without having to waste prod-realm public IP addresses. We should investigate if we can move the wiki replicas behind that, or with some other mechanism, like having the current dbproxy hosts advertise VIPs to cloud-private.
Description
Details
Status | Subtype | Assigned | Task | ||
---|---|---|---|---|---|
Resolved | aborrero | T296411 cloud: decide on general idea for having cloud-dedicated hardware provide service in the cloud realm & the internet | |||
Resolved | aborrero | T297596 have cloud hardware servers in the cloud realm using a dedicated LB layer | |||
Open | None | T297026 Automate maintain-views workflow | |||
Open | taavi | T300427 Automate maintain-views replica depooling | |||
Resolved | taavi | T355115 Remove cloud-support1-c-eqiad VLAN | |||
Resolved | taavi | T346947 Move wiki replicas behind cloudlb | |||
Resolved | taavi | T351087 Migrate cloudlb hosts to nftables | |||
Resolved | jbond | T351094 nftables ignores drange filter for IPv6 if drange only has IPv4 addresses |
Event Timeline
Change 973769 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/homer/public@master] cr-labs: permit cloudlb to wiki replicas
Change 973761 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/puppet@production] Add wiki replicas to cloudlb
Change 973777 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/puppet@production] P:wmcs: wikireplicas: allow access from cloudlb
Change 973769 merged by jenkins-bot:
[operations/homer/public@master] cr-labs: permit cloudlb to wiki replicas
Change 974534 had a related patch set uploaded (by Cathal Mooney; author: Cathal Mooney):
[operations/dns@master] Remove includes for subnets from cloud-support1-a-eqiad
Change 974534 merged by Cathal Mooney:
[operations/dns@master] Remove includes for subnets from cloud-support1-a-eqiad
Change 973777 merged by Majavah:
[operations/puppet@production] P:wmcs: wikireplicas: allow access from cloudlb
Change 976688 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/puppet@production] openstack: update wiki replica DNS to cloudlb
Change 973761 merged by Majavah:
[operations/puppet@production] Add wiki replicas to cloudlb
Change 976735 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/puppet@production] P:etcd: generate wiki replica pool accounts
Change 977081 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/alerts@master] team-wmcs: Adapt cloudlb alerts for wiki replicas
Change 977081 merged by jenkins-bot:
[operations/alerts@master] team-wmcs: Adapt cloudlb alerts for wiki replicas
Change 976688 merged by Majavah:
[operations/puppet@production] openstack: update wiki replica DNS to cloudlb
Mentioned in SAL (#wikimedia-cloud) [2023-11-28T14:18:07Z] <taavi> moving wiki replica DNS to use cloudlbs instead of the old proxy VMs T346947
Change 978539 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/puppet@production] hieradata: unconfigure wiki replica LVS services
Change 979045 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/puppet@production] cloudlb: wikireplicas: fix timeouts
Change 979045 merged by Majavah:
[operations/puppet@production] cloudlb: wikireplicas: fix timeouts
Mentioned in SAL (#wikimedia-cloud) [2024-01-08T12:57:11Z] <taavi> shut off old wiki replica proxies in cloud vps T346947
Change 988483 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/puppet@production] hieradata: remove wikireplica service catalog entries
Change 988484 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/dns@master] wmnet: remove aliases for dbproxy1018/9
Change 988484 merged by Majavah:
[operations/dns@master] wmnet: remove aliases for dbproxy1018/9
Change 978539 merged by Majavah:
[operations/puppet@production] hieradata: unconfigure wiki replica LVS services
Mentioned in SAL (#wikimedia-operations) [2024-01-08T16:09:13Z] <taavi> restart pybal on lvs1020 - T346947
Mentioned in SAL (#wikimedia-operations) [2024-01-08T16:15:27Z] <taavi> restart pybal on lvs1018 - T346947
Mentioned in SAL (#wikimedia-operations) [2024-01-08T16:20:54Z] <taavi> lvs1020: sudo ipvsadm --delete-service --tcp-service 208.80.154.242:3311 (and all the way to :3318) - T346947
Mentioned in SAL (#wikimedia-operations) [2024-01-08T16:21:28Z] <taavi> lvs1020: sudo ipvsadm --delete-service --tcp-service 208.80.154.243:3311 (and all the way to :3318) - T346947
Mentioned in SAL (#wikimedia-operations) [2024-01-08T16:23:17Z] <taavi> lvs1018: sudo ipvsadm --delete-service --tcp-service 208.80.154.242:3311 (and all the way to :3318) - T346947
Mentioned in SAL (#wikimedia-operations) [2024-01-08T16:24:18Z] <taavi> lvs1018: sudo ipvsadm --delete-service --tcp-service 208.80.154.243:3311 (and all the way to :3318) - T346947
Change 988483 merged by Majavah:
[operations/puppet@production] hieradata: remove wikireplica service catalog entries
Change 988670 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/puppet@production] O:mariadb::proxy: remove LVS realserver profile
Change 988670 merged by Majavah:
[operations/puppet@production] O:mariadb::proxy::replicas: remove LVS realserver profile
Change 988681 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/puppet@production] Move dbproxy1018/9 to insetup
Change 989087 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/puppet@production] conftool-data: Remove wiki replica dbproxies
Change 989088 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/puppet@production] mariadb: remove grants and firewall rules for dbproxy1018/9
Change 989087 merged by Majavah:
[operations/puppet@production] conftool-data: Remove wiki replica dbproxies
Mentioned in SAL (#wikimedia-cloud) [2024-01-09T12:01:07Z] <taavi> delete old neutron ports used for old wiki replica VIPs T346947
Change 988681 merged by Majavah:
[operations/puppet@production] Move dbproxy1018/9 to insetup
cookbooks.sre.hosts.decommission executed by taavi@cumin1002 for hosts: dbproxy[1018-1019].eqiad.wmnet
- dbproxy1018.eqiad.wmnet (PASS)
- Downtimed host on Icinga/Alertmanager
- Found physical host
- Downtimed management interface on Alertmanager
- Wiped all swraid, partition-table and filesystem signatures
- Powered off
- [Netbox] Set status to Decommissioning, deleted all non-mgmt IPs, updated switch interfaces (disabled, removed vlans, etc)
- Configured the linked switch interface(s)
- Removed from DebMonitor
- Removed from Puppet master and PuppetDB
- dbproxy1019.eqiad.wmnet (PASS)
- Downtimed host on Icinga/Alertmanager
- Found physical host
- Downtimed management interface on Alertmanager
- Wiped all swraid, partition-table and filesystem signatures
- Powered off
- [Netbox] Set status to Decommissioning, deleted all non-mgmt IPs, updated switch interfaces (disabled, removed vlans, etc)
- Configured the linked switch interface(s)
- Removed from DebMonitor
- Removed from Puppet master and PuppetDB
Change 989541 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/puppet@production] site: remove dbproxy1018/9
Change 989541 merged by Majavah:
[operations/puppet@production] site: remove dbproxy1018/9
Change 989088 merged by Majavah:
[operations/puppet@production] mariadb: remove grants and firewall rules for dbproxy1018/9
Mentioned in SAL (#wikimedia-cloud) [2024-01-16T09:41:22Z] <taavi> drop dbproxy1018/9 grants from all clouddb hosts T346947
Change 990957 had a related patch set uploaded (by Majavah; author: Majavah):
[operations/software@master] report_users: drop dbproxy1018/9
Change 990957 merged by jenkins-bot:
[operations/software@master] report_users: drop dbproxy1018/9