and Consume Consume Private Cloud Public Cloud Public Cloud VMware Cloud Native Apps Portfolio PKS VKE VIC Beta 5 ©2018 VMware, Inc. vSphere Integrated Containers (VIC) 6 ©2018 VMware, Inc. vSphere support Docker API Resource Pool 75 Ghz, 768GB 8 ©2018 VMware, Inc. Pivotal Container Service (PKS) 9 ©2018 VMware, Inc. Kubernetes Is Only One Layer of the Container Service Stack Image Registry Suite NSX-T BOSH (K8s LCM) BOSH (cluster LCM) Kubernetes Harbor P K S 10 ©2018 VMware, Inc. PKS on VMware SDDC Rapidly deliver and operationalize next-generation apps Physical Infrastructure

e.g. com.example.something • timeout=[millis]: timeout in milliseconds • create-pks: create primary keys • drop-pks: drop primary keys • create-constraints: create indexes and foreign keys after loading

e.g. com.example.something • timeout=[millis]: timeout in milliseconds • create-pks: create primary keys • drop-pks: drop primary keys • create-constraints: create indexes and foreign keys after loading

only pass events for pks 1 to 10 T3: select * from t1 where pk > 10 limit 10 GTID: 112, Last PK 20 send rows to target Another gotcha!: Note that at T3 when we selected the pks from 11 to 20 the gtid from T3 we would still have an inconsistent state, if transactions 111 and 112 affected the rows from pks 1 to 10. This leads us to the following flow: T1: select * from t1 where pk > 0 limit 10. GTID: 100 only pass events for pks 1 to 10 T3: select * from t1 where pk > 10 limit 10 GTID: 112, Last PK 20 T4: replicate from 111 to 112 only pass events for pks 1 to 10 T5: Send rows for pks 11 to 20 to target

only pass events for pks 1 to 10 T3: select * from t1 where pk > 10 limit 10 GTID: 112, Last PK 20 send rows to target Another gotcha!: Note that at T3 when we selected the pks from 11 to 20 the gtid from T3 we would still have an inconsistent state, if transactions 111 and 112 affected the rows from pks 1 to 10. This leads us to the following flow: T1: select * from t1 where pk > 0 limit 10. GTID: 100 only pass events for pks 1 to 10 T3: select * from t1 where pk > 10 limit 10 GTID: 112, Last PK 20 T4: replicate from 111 to 112 only pass events for pks 1 to 10 T5: Send rows for pks 11 to 20 to target

only pass events for pks 1 to 10 T3: select * from t1 where pk > 10 limit 10 GTID: 112, Last PK 20 send rows to target Another gotcha!: Note that at T3 when we selected the pks from 11 to 20 the gtid from T3 we would still have an inconsistent state, if transactions 111 and 112 affected the rows from pks 1 to 10. This leads us to the following flow: T1: select * from t1 where pk > 0 limit 10. GTID: 100 only pass events for pks 1 to 10 T3: select * from t1 where pk > 10 limit 10 GTID: 112, Last PK 20 T4: replicate from 111 to 112 only pass events for pks 1 to 10 T5: Send rows for pks 11 to 20 to target

only pass events for pks 1 to 10 T3: select * from t1 where pk > 10 limit 10 GTID: 112, Last PK 20 send rows to target Another gotcha!: Note that at T3 when we selected the pks from 11 to 20 the gtid from T3 we would still have an inconsistent state, if transactions 111 and 112 affected the rows from pks 1 to 10. This leads us to the following flow: T1: select * from t1 where pk > 0 limit 10. GTID: 100 only pass events for pks 1 to 10 T3: select * from t1 where pk > 10 limit 10 GTID: 112, Last PK 20 T4: replicate from 111 to 112 only pass events for pks 1 to 10 T5: Send rows for pks 11 to 20 to target

only pass events for pks 1 to 10 T3: select * from t1 where pk > 10 limit 10 GTID: 112, Last PK 20 send rows to target Another gotcha!: Note that at T3 when we selected the pks from 11 to 20 the gtid from T3 we would still have an inconsistent state, if transactions 111 and 112 affected the rows from pks 1 to 10. This leads us to the following flow: T1: select * from t1 where pk > 0 limit 10. GTID: 100 only pass events for pks 1 to 10 T3: select * from t1 where pk > 10 limit 10 GTID: 112, Last PK 20 T4: replicate from 111 to 112 only pass events for pks 1 to 10 T5: Send rows for pks 11 to 20 to target

by VMware China, maintained by whole community Integrated into enterprise products: VIC and PKS GitHub Repo: https://github.com/go harbor/harbor/ Apache 2.0 license An open source trusted cloud

users with a vSphere or VMware Cloud deployment. The origins of TKG are Pivotal Container Service (PKS), whose origins, in turn, are from Cloud Foundry, a platform originally developed and spun off by