Terminology and definitions
Terms used when working with Db2® Mirror are defined in this section.
- ACTIVE replication state
- The replication state when Db2 Mirror is running normally. Db2 Mirror is keeping replication-eligible objects synchronized.
- BLOCKED replication state
- The replication state on the secondary node when active replication is suspended. Changes to most replicated objects will be blocked until active replication is resumed.
- Client File System (CFS)
- The integrated file system (IFS) Mutable File System instance on the node where the independent auxiliary storage pool (IASP) is not varied on. The CFS sends IFS requests to the corresponding Server File System node.
- Clone
- The initial action during configuration and setup that results in the two nodes being identical. A clone is also performed when defining a database IASP.
- Cluster
- A set of IBM® i partitions called cluster nodes. Cluster technology provides the critical infrastructure to enable high availability solutions. It provides mechanisms for defining resilient resources in an environment, detecting outages, and responding to these failures.
- Cluster monitor
- Cluster monitors are used for advanced node failure detection when communication between cluster nodes has failed unexpectedly. Cluster monitors provide the capability to query the Hardware Management Console (HMC) for the status of other cluster nodes.
- Cluster node
- An IBM i partition that is a member of a cluster. Cluster nodes can share resources that are required to be highly available for your business with other nodes in the cluster. This allows other nodes in the cluster to provide backup in the case of an outage.
- Cluster resource group (CRG)
- An entity within a cluster that provides the mechanisms to monitor and manage the switching of resources between cluster nodes.
- Cold cloning
- The setup source node is powered off during the cloning process. This ensures the most accurate copy of all data on the source node because no changes are being made while the copy takes place.
Copy node
- The node that is modified by the software reclone process. Replicated objects on this node are
replaced with versions from the source node. Contrast with source node.
- Database IASP
- An independent auxiliary storage pool (IASP) that contains database and non-database objects eligible for replication by Db2 Mirror.
- Default inclusion state
- A value that indicates the inclusion state for an object when the Replication Criteria List has no defined rule that applies to the object. A default inclusion state is defined for SYSBAS and for each database IASP.
- Encrypted RoCE v2
- Encrypted RDMA over Converged Ethernet (RoCE) version 2 is an encrypted and routable
RDMA protocol. RoCE version 2 with IPsec encryption is an internet layer protocol which enables
packets to be routed securely by using IPsec protocols to provide data authentication, integrity,
and confidentiality. The ports of RoCE adapters can be connected through multiple switches or
routers, up to a maximum distance of 10 kilometers. A VPN connection must be configured for each
RDMA link endpoint with data transforms (encryption protocols) supported by the adapter. A VPN
connection must be available for the RDMA link to start. If the VPN connection is ended, the RDMA
link is also ended. Data is encrypted using IPsec and transferred using the UDP protocol and UDP
destination port number 4791. Contrast with RoCE v1 or RoCE v2.
- Hardware reclone
- A type of reclone operation where hardware replication, either FlashCopy® or storage copy services technology, is used to make the two nodes or database IASPs identical again. Contrast with software reclone.
- Health monitor
- A system job named QMRDBHLTHM that is used to continuously monitor for any issues that might arise relating to the Db2 Mirror environment and act when an issue is detected.
- IFS IASP
- An independent auxiliary storage pool (IASP) that contains IFS objects. The IFS objects within the IASP are accessible from both nodes.
- Inclusion state
- The replication action associated with an object. The default inclusion state along with rules in the Replication Criteria List determine whether an object should be included or excluded from replication.
- Job Queue List (JQL)
- A list of job queues that are being tracked by Submitted Job Tracker. On each node where
job queues are being tracked, there is a database file named QSYS/QAMRDJQL used to store the JQL.
The JQL is not replicated in a Db2 Mirror environment,
therefore must be configured on each node separately.
- Job tracking file
- A database file that keeps track of job submit parameters and job status for any jobs being
tracked by Submitted Job Tracker. In a Db2 Mirror
environment, a pair of job tracking files is populated and replicated for SYSBAS and for each
database IASP. One job tracking file, named QSBMJOBTRK/QAMRDJTS1, is used to store information about tracked jobs
submitted on one node and it is replicated to the second node. A second job tracking file, named
QSBMJOBTRK/QAMRDJTT2,
is used to store information about tracked jobs submitted on the second node and it is replicated to
the first node.
- Mirror Version List (MVL)
- The list that contains information about registered version information entries for a specific
feature or function.
- Mutable File System (MFS)
- An IFS file system type that is used by Db2 Mirror to provide switchable file systems for switchable IFS IASPs in the Db2 Mirror environment.
- Network Redundancy Group (NRG)
- A collection of one or more physical Remote Direct Memory Access (RDMA) enabled links to a target system. NRGs provide automated physical network redundancy, failover, and load balancing for Db2 Mirror connections between the nodes. The five system defined NRGs for Db2 Mirror are: MIRROR_DATABASE, MIRROR_ENGINE, MIRROR_IFS, MIRROR_OTHER, and MIRROR_RESYNC.
- Network Time Protocol (NTP) server
- A server used to automatically synchronize the system clocks of the primary node and secondary node.
- Node
- A partition used by Db2 Mirror.
- Non-routable RDMA
- Communicating RDMA hosts must be in the same Ethernet broadcast domain and in the same subnet
because an internet layer protocol is not used. Contrast with routable RDMA.
- NOT MIRRORED replication state
- The replication state when Db2 Mirror is not installed or has not completed initial configuration.
- Object Tracking List (OTL)
- The list that contains information about object operations that need to be replicated between the nodes and information about objects that have already been resynchronized. The OTL is also used to perform some active replication operations; for example, when a rule is added to the Replication Criteria List that requires an object to be replicated to the other node.
- Primary node
- The node that will enter TRACKING state when active replication is suspended between the nodes. Contrast with secondary node.
- Pseudo object type
- A special object type used in the Replication Criteria List and in the Object Tracking List to reference an entity that is not a system object. Pseudo object types are defined for system level environment variables (*ENVVAR), security attributes (*SECATR), system values (*SYSVAL), and function usage identifiers (*FCNUSG).
- Quorum data
- Cluster-wide data that is stored on all the nodes in the cluster. It is used by Db2 Mirror to verify status and information about the nodes when they are unable to communicate.
- Quorum node
- A third node added to the cluster to ensure the reliability of quorum data.
- Remote Direct Memory Access (RDMA) over Converged Ethernet (RoCE)
- The communication protocol that is required by Db2 Mirror to provide a fast network connection with low latency and high bandwidth between the two nodes.
- Reclone
- A clone that is performed after Db2 Mirror has already been operational.
- Replication Criteria List (RCL)
- A rules engine used to determine whether objects are included or excluded from replication.
- Replication criteria rule
- A rule in the Replication Criteria List that defines the inclusion state for an object or a group of objects. Each rule contains criteria for IASP name or SYSBAS, library name, object type, and object name.
- Replication state
- The current state of replication. The states can be different for SYSBAS and for each database IASP.
- Resynchronization
- The replication of changes between the primary node and the secondary node when replication is resumed after being in the TRACKING state and BLOCKED state. When phase one of resynchronization is complete and resynchronization has obtained necessary locks for the objects to be replicated, the replication state will become ACTIVE.
- RoCE v1
- RDMA over Converged Ethernet (RoCE) version 1 is a non-encrypted and non-routable
RDMA protocol. It transfers data using the Ethernet link layer protocol with Ethertype 0x8915
which allows communication between two hosts within the same Ethernet broadcast domain. The ports of
RoCE adapters can be cabled directly together or connected by two cables and a single switch, which
provides a maximum distance of 200 meters. Adapter FC EC78 and LR Transceivers (NVIDIA MMA1L10-CR
Optical Transceiver 100GbE QSFP28 LC-LC LR4) can support up to 10 kilometers. Communicating host IP
addresses must be in the same Ethernet broadcast domain and in the same subnet. This type of link is
available only if non-encrypted RDMA is allowed in your environment. Contrast with RoCE v2 or
encrypted RoCE v2.
- RoCE v2
- RDMA over Converged Ethernet (RoCE) version 2 is a non-encrypted and routable RDMA
protocol. It is an internet layer protocol which enables packets to be routed. The ports of RoCE
adapters can be connected through multiple switches or routers, up to a maximum distance of 10
kilometers. Data is not encrypted and is transferred using the UDP protocol and UDP destination port
number 4791. This type of link is available only if non-encrypted RDMA is allowed in your
environment. Contrast with RoCE v1 or encrypted RoCE v2.
- Routable RDMA
- Communicating RDMA hosts can be in separate networks because an internet layer protocol is used
which enables packets to be routed through IP networks. The NRG link IP addresses can be in the same
or different Ethernet broadcast domain and subnet. If in different subnets, each link needs a TCP/IP
route. Contrast with non-routable RDMA.
- Secondary node
- The node that will enter BLOCKED state when active replication is suspended between the nodes. Contrast with primary node.
- Server File System (SFS)
- The integrated file system (IFS) Mutable File System instance on the node where the independent auxiliary storage pool (IASP) is varied on. The SFS node is where IFS requests against the IASP are processed. Requests initiated from the Client File System (CFS) are received, executed, and the appropriate response is sent back to the CFS.
- Setup copy node
- The node that is created by the setup process. It is created by cloning the setup source node and then configured by Db2 Mirror automatically during its first IPL. After the setup process is complete, this will be the secondary node. Contrast with setup source node.
- Setup source node
- An IBM i partition that is often an already established production partition where your applications are running. The setup process will make an identical copy of this node and establish the Db2 Mirror environment. After the setup process is complete, this will be the primary node. Contrast with setup copy node.
- Software reclone
- A type of reclone operation where only the replicated objects are copied from the source of the
reclone to the target. Objects which are dependent on those replicated objects may be affected on
the target node, but other non-replicated data, as well as the OS and products on the target will
not be affected. A software reclone can be done on SYSBAS or on a database IASP. Contrast with
hardware reclone.
- Source node
-
- The node that initiated the replication-eligible operation. Contrast with target node.
- The node that initiated the software reclone. Replicated objects from this node
are saved and restored to the copy node. Contrast with copy node.
- Submitted Job Tracker
- An IBM i work management feature that can be used in the
Db2 Mirror environment to capture information about jobs as
they are submitted to job queues on each node. The Job Queue List is used to manage which job
queues are being tracked. Tracked job information is stored in a job tracking file.
The information in the job tracking file can help prepare for maintenance, recover from a
node failure, or balance workload. The information can be used to resubmit the jobs on the original
node where the job was tracked or another node.
- Takeover IP address
- An IP address which is part of a takeover IP address group.
- Takeover IP address group
- A group of IP addresses which can be switched between Db2 Mirror nodes as one entity.
- Target node
- The node that did not initiate the replication-eligible operation. Contrast with source node.
- Tracked job
- A job submitted to a job queue whose job parameters and job status are being tracked by
Submitted Job Tracker in a job tracking file.
- TRACKING replication state
- The replication state on the primary node when active replication is suspended. Db2 Mirror tracks all replication-eligible operations in the Object Tracking List.
- Version information
- Registered version information in the Mirror Version List is used to determine the common
active version that is used by both nodes to coordinate specific features or functions.
- Warm cloning
- The setup source node is not powered off during the cloning process. This allows changes to occur on the setup source node during the cloning process. The changes to replicated objects will be tracked by Db2 Mirror. Following the clone, the tracked changes will be synchronized to make the replicated objects consistent between the nodes.
1 For each database IASP,
the QAMRDJTS file exists in library QSJTnnnnn.
2 For each database IASP, the QAMRDJTT file exists in library QSJTnnnnn.