logo

ALM GLOSSARY

A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z

Welcome to the Glossary for all our SCM/ALM products.

For more detailed information on each product please visit the specific user guide:
Access Control Plus
Gerrit MultiSite
Git MultiSite
Subversion MultiSite Plus

If you use also use our Big Data products, see their glossary here.

A

Acceptor

This is a Paxos term which identifies a node that can vote on a proposal. In Content Distribution an acceptor is the recipient node for an incoming Agreement message initiated by a Proposer.

Access Control

An legacy product which, due to it being implemented as a proxy, enabled more granular authorization for Subversion than is possible by Apache alone.

Access Control Plus (ACP)

Product which provides a single solution to fully protect valuable intellectual property across widely distributed Subversion and Git repositories.

Active node

A node in a replication group which has been enabled to request changes be made to a set of repositories. It may or may not be a voter. If it is not a voter then it does not play a role in scheduling the changes it requests.

Active voter node

An Active node which also votes on the order in which transactions are played out.

Admin Console

Provides administrators with a UI based mechanism for controlling the application.

Agreement

Each step that the DConE replicated state machine executes is called an agreement.

Agreement Manager

Individual agreement steps of the replicated state machine are executed under the purview of one or more Agreement Managers.

API

Application program interface.

Authentication

Verifies the credentials presented - essentially proving you are who you say you are.

Authorization

Defines, with rules, what users are allowed to do on a system.

AuthN

See Authentication

AuthZ

See Authorization

B

Batched Replication

Objects changed in a batch of user transactions are replicated in a single replication transaction. Irrespective of whether a replication is transactional or batch based only the part of the transaction that is of interest to that site is applied at the receiving site. Batched transactions can utilize the self-healing features. For example, the network connection can be disabled between the distinguished node site and the other sites on a periodic basis (e.g. twice a day). When the network is reconnected, the remote replicator will reach out to the distinguished node to pick up all of the commits that were done at the distinguished node site during the network outage.
Only relevant to SVN MultiSite 4.x or earlier.

Bypassable Transaction

A transaction is bypassable if all nodes in Replication Group fail on this exact transaction.
Only relevant to SVN MultiSite 4.x or earlier.

C

CA

See Certificate Authority

CBT

See Change Block Tracking.

C-Git

The implementation of Git written in C. This is not to be confused with cgit, a fast web interface for the git version control system. For GitMS you need to install the matching binaries for your version of GitMS.

CentOS

A community-supported, free and open source operating system based on Red Hat Enterprise Linux. It exists to provide a free enterprise class computing platform and strives to maintain 100% binary compatibility with its upstream distribution. CentOS stands for Community ENTerprise Operating System.

Certificate Authority (CA)

A body which issues digital certificates.
A certificate is a digitally signed statement from one entity (the issuer), saying that the public key (and some other information) of another entity (the subject) has some specific value. These trusted certificates can then be used to create secure connections to a server.

Change Block Tracking (CBT)

A technology which allows the incremental backup of VMs. This means that only the blocks which have changed are backed up, rather than every block. If CBT is enabled make sure to account for it, for example when restoring repositories.

Clustering

Clustering products are implemented as a transparent gateway between the clients and the servers, and acts as a network proxies for the local server. There is a one-to-one relationship between replicator and SVN Server. With clustering, a load balancer directs a user to one node in the cluster but clients write activity must be replicated from one node to the others in the cluster.

D

Data Transfer Object (DTO)

An object which transfers data between software application subsystems.

DConE

Distributed Coordinated Engine, the software engine underlying replication. Read the Whitepaper on our DConE technology.

Deterministic State Machine (DSM)

An object whose principle job is obtaining agreements on the ordering of proposals as part of the DConE engine. Each DSM has a membership of nodes. Each node can have one or more of the Paxos roles (Proposer, Acceptor, Learner).

Distinguished Name (DN)

Used for unique identification in LDAP.

Distinguished Node

A node whose vote is slightly more than 1 (also known as a Tiebreaker).

DN

See Distinguished Name.

DSM

See Deterministic State Machine.

DTO

See Data Transfer Object.

E

Ecosystem

The combined set of nodes that have been inducted together.

Emergency Reconfiguration (EMR)

A tool which allows one or more lost nodes to be removed from membership, allowing the remaining nodes to continue replicating using a new membership. This is a permanent change to the replication system which cannot be undone. EMR is only necessary if the remaining nodes are missing quorum from 1 or more voting memberships.

EMR

See Emergency Reconfiguration.

F

Flume

Flume is an Apache product for efficiently collecting, aggregating and moving large amounts of log data.

FQDN

See Fully Qualified Domain Name.

FSFS

A Subversion filesystem implementation.

Fully Qualified Domain Name (FQDN)

A domain name that specifies its exact location within the tree hierarchy of the DNS. It is unambiguous.

G

Garbage Collection

A type of automatic memory management. It allows system resources to be freed up by removing objects that are no longer required.

Gerrit

An open source code review tool which integrates closely with Git. It was named after the Dutch designer Gerrit Rietveld.

Gerrit MultiSite (GerritMS)

Allows Gerrit events to be replicated to and from any location with no single point of failure.

Git

An open source version control system.

Git MultiSite (GitMS)

Allows Git repositories to be replicated to and from any location with no single point of failure and no downtime.

Gitolite

An access control layer on top of Git which can be integrated with Git MultiSite.

Global Sequence Number (GSN)

The ordered number of an agreement associated with a DSM.

Globally Unique Identifier (GUID)

Each node is assigned a GUID when added to a Replication Group. The nodes identify each other by their GUIDs.

Group Evolution

A feature that allows changes to a previously defined Replication Group.

Growl

Growl messages provide immediate feedback in response to a user’s interactions with the Admin UI. They appear in the top right-hand corner of the screen and persist for a brief period or until the screen is refreshed or changed.

GSN

See Global Sequence Number

GUID

See Globally Unique Identifier

H

Helper Node

A node that is used as part of a repository repair procedure. The helper node is suspended from receiving updates to the repositories in the replication group so that a copy of the repositories can be transferred (rsync’ed etc) to the damaged node.

High Availability

Provides continuous hot backup, while making failover and disaster recovery automatic and transparent for both developers and administrators.

Hooks

Scripts triggered by specific repository events, they allow administrators to have more control over their repository environment.

I

Inductee

A node yet to be inducted.

Install node

The node where you run the product install program. This is the first node in your ecosystem. This term is only relevant to SVN MultiSite 4.x or earlier.

J

JGit

The Java library from Eclipse. This is installed as part of GitMS.

JIRA

Browser-based bug, issue, task and defect tracking system, and project management software solution used for open source and enterprise projects. JIRA isn’t an acronym, it’s a contraction of 'Gojira'. When listing product features and bug fixes we’ll usually add a reference code for the corresponding JIRA issue.

Jenkins

An open source automation server.

K

KB

See Knowledgebase.

Kerberos

A network authentication system.

Knowledgebase (KB)

The Knowledgebase contains updates and product information.

L

LDAP

See Lightweight Directory Access Protocol.

Learner

This is a Paxos term that identifies any node that wants to learn the outcome of an Agreement step. In Content Distribution a learner is any node that has a repository and needs to have agreements delivered and executed.

Lightweight Directory Access Protocol (LDAP)

An Internet protocol that enables client programs to access distributed directory services.

Lock

A mechanism to enforce access limits. Locks need to be the same on all connected nodes.

M

Master node

In the context of node distinction, master is a legacy term used to describe the "primary" instance. The "primary" instance is "read-write" while the "secondary" instance is "read-only".

Master-slave

In the context of a replication model, a legacy term used to describe "primary-secondary" replication.

Membership

A term that is used to describe sets of nodes that are connected to each other as part of a replication group (or other DSM).

N

Node

A server on which a replicator product (e.g. Git MultiSite) is installed.

O

One copy equivalence

All replicas are functionally equivalent copies of each other. The theory and practice is that if two or more nodes start with identical copies of data and all changes to that data are applied in the same order to each copy, at the end of X transactions, the copies will all remain equivalent.

Owl

A fan of Sheffield Wednesday Football Club.

P

Passive node

A node on which repositories receive updates from other nodes, but does not permit any modifications to be requested locally (cannot propose changes). This node type is currently deprecated and should no longer be used.

Passive voter node

A passive node which also takes part in the vote for transaction ordering agreement but cannot propose any changes.

prefs.xml

The preferences file is used in SVN MultiSite 4.2 or earlier (or the old CVS MultiSite product). prefs.xml contains configuration information and is specific to each node. The file is located in svn-replicator/config (or cvs-replicator/config). See the knowledgebase article Prefs.xml Basic Syntax.

Primary node

In the context of node distinction, the "primary" instance is "read-write" while the "secondary" instance is "read-only".

Primary-secondary

Similar to the legacy term "master-slave". Primary-secondary replication indicates the "primary" instance is "read-write" while the "secondary" instance is "read-only".

Proposer

This is a Paxos term which identifies a node which can propose changes. In Content Distribution, as well as proposing changes the proposer is also involved in conflict resolution as part of the voting process. Resolution comes from interactions between proposers and accepters and the losers are expect to re-propose.

Q

Quarantined Mode

When activating a membership or Replication Group, new nodes start in quarantined mode. This means they will not participate in replication until they gain membership. This is to allow for testing of the new node’s SCM configuration before executing replicating transactions. Only relevant to SVN MultiSite 4.x or earlier.

Quorum

Any subset of a membership that can reach agreement to determine transaction order. In the case of an even number of nodes, the distinguished node (tiebreaker) settles a conflict.

Read more about SVN MultiSite Plus’s more flexible approach to building quorums - Replication Strategy.

Quorum (Singleton)

Singleton Quorum - only one of the nodes in the membership is an acceptor. The Singleton quorum offers the fastest response time for those users working at the distinguished node, because as soon as the distinguished node determines that a transaction can be processed in the correct order, the transaction is sent to SVN. Any replicator except the distinguished node can go down, but the replication group continues. The replication group replays the missing transactions when that node rejoins the group. However, the Singleton quorum also represents a single point of failure, since replication halts if the distinguished node fails (e.g. all repository change requests fail).

Quorum (Majority)

Majority Quorum is another quorum option, whereby you specify that a majority of the sites must agree on transaction order before any transaction is committed. Having a majority quorum ensures that if one site goes down in a replication group, even the distinguished node, the other sites can continue uninterrupted, as long as a majority of the sites remain available. Self-healing will take place for any node that is down while the rest of the membership is making changes.

For example, in a four node replication group, three sites make the quorum (three sites must agree about transaction ordering). If two nodes want one transaction first, and the other two want another transaction first, then the distinguished node gets a weighted vote. The group with the distinguished node determines the transaction ordering. With an even number of nodes with majority quorum, you can schedule the distinguished node to rotate to different nodes around the world, to 'follow the sun'.

Only relevant to SVN MultiSite 4.x or earlier.

R

Read-only mode

As a result of a failed write operation a DSM (or node in the MS 4.x product or earlier) can go into read-only mode. Write operations cannot proceed until the problematic transaction has been fixed. This can be Global (on all nodes) or local (just for users at this node).

Recovery Point Objective (RPO)

In disaster recovery the RPO is the time before a system failure at which the recovered system will restart. An RPO of 2 hours means that, in the event of a failure, the system would be brought back to the point that is no more than 2 hours before the failure occurred.

Recovery Time Objective (RTO)

In disaster recovery the RTO is the time in which the system should be recovered from a failure. With an RTO of 2 hours, the system should be recovered no more than two hours after the failure occurred. High Availability offers an RTO approaching 0, meaning that recovery is nearly immediate.

RO

See Read only

Replica

An SVN/Git repository that is an exact equivalent or copy of another SVN/Git repository.

Replication Group

A collection of repositories that share a membership and are therefore all present on the same set of nodes.

Replicator

A replicator is synonymous with an installed instance of the product or "node".

Repository

A collection of files and directories, grouped together in a specific database which also records a complete history of all the changes that have ever been made to these files.

Resources

In Access Control Plus, resources are repositories that have been activated by being associated with a repository template.

RESTful API

Representational State Transfer Application Program Interface.

RPO

See Recovery Point Objective.

Rsync

A file copying tool.

RTO

See Recovery Time Objective.

S

SCM

See Software Configuration Management

Secondary node

In the context of node distinction, the "secondary" instance is "read-only" while the "primary" instance is "read-write".

SHA1

Secure Hash Algorithm 1. This is a 160-bit hash value normally rendered as a 40 digit hexadecimal number. Although originally developed for security purposes, SHA1 is used in Git to ensure data hasn’t changed, for example due to accidental corruption.

Sideline

A sidelined repository replica is a replica that will no longer get updates from other replicas in the family. The other replicas will not preserve operations for a sidelined replica thereby preventing them from running out of memory. Replicas become sidelined if they are out of communication for an extended amount of time and the number of outstanding agreements exceeds a tunable maximum. Sidelined replicas can be brought back into normal operation via a repair procedure.

Site

A physical location containing servers or virtual machines where one or more replication products are installed.

Software Configuration Management (SCM)

The field dedicated to tracking and controlling changes in the software. Products that support two popular SCM solutions: SVN and Git.

SSH

Secure Shell (SSH) is a means of getting secure access to a remote computer. It can be used for authentication.

SSL

Secure Socket Layer (SSL) is a commonly used encryption protocol.

Subversion (SVN)

An open source version control system.

SVN

See Subversion.

SVN MultiSite Plus (SVN MSP or MSP)

Product which allows SVN events to be replicated to and from any location with no single point of failure and no downtime.

Synchronized Stop

A special transaction that will prevent further write transactions from happening. A replicator will process transactions normally up to this special transaction and the node will enter a "stopping" state. This causes all nodes to stop after processing up through the exact same GSN. A Synchronized Stop is typically actioned prior to administrative tasks such as a product upgrade. This transaction requires Unanimous Agreement in order to complete. Even if one of the nodes is not available, the other nodes will enter the stopping state and prevent write transactions. Unanimous Agreement means that all nodes need to be available for a synchronized stop to complete.

Systemd

For up to date information see the GitHub page for systemd news.

T

Talkback

A bash script that is provided in your product installation to collect product meta-data in case you need to report an issue to the Support team.

Tiebreaker node

A node which gets to cast the determining vote if there are an even number of voters in the Replication Group. The Tiebreaker can be applied any type of voter node. This term is specific to 4.x and earlier.
In MSP, the tiebreaker role is assigned on a per-Replication Group basis and is not one specific node.

U

UI

User interface. The console with which users interact.

Universally unique identifier (UUID)

A 128-bit identifier that is created in such a way that no other identifier will ever be its equal (see GUID).

UUID

See Universally unique identifier.

V

Voter only node

A Voter only node doesn’t store any repository data, it’s only purpose is to cast a vote on transaction ordering. Voter only nodes are currently deprecated and should no longer be used.

W

WebDAV

Web-based Distributed Authoring and Versioning, or WebDAV is a set of extensions to the Hypertext Transfer Protocol (HTTP) that allows computer-users to edit and manage files collaboratively on remote web servers.

Wildcard

A symbol used to replace or represent one or more characters.

X

XML

Extensible Markup Language.

Y

Yeturu Aahlad

Dr Aahlad, chief scientist and is recognized as a authority on distributed computing in which he currently holds 3 patents.

Z

ZooKeeper

A software project of Apache which is not relevant to these ALM products but is a word which begins with Z (and there aren’t many). It is however relevant to our Big Data products.