Knowledge Center

103 - How a Request Flows Through the System

Updated on

How a Request Flows Through the System

The sequence every archive, restore and delete follows, and the service that drives it.

Three components do the work: the DMSync database, the GDSI.DMSyncArchiveService Windows service, and the archive storage target.

The service

GDSI.DMSyncArchiveService polls the rules tables every 20 to 25 seconds and processes the request queue. Everything automated flows through it. There is no scheduled task to configure — the service runs continuously.

The sequence

  1. A request is raised, either by a rule the service picked up or by a user action in Work 10.
  2. A row is inserted into ArchiveRequests describing the operation and its scope.
  3. The service processes each object in the request, writing to or reading from the storage target.
  4. One ArchiveAudit row is written per document processed.
  5. The affected Documents and Workspaces rows are stamped or cleared.
  6. The request row is updated with Status, SuccessProcessedCount and ProcessedAt.

Telling the two origins apart

Every request records where it came from. This is the fastest way to tell whether a user or a rule caused something.

Column
Rule-driven

From Work 10

CreatedBy
SchedulerThe acting account
RuleNameThe rule nameNULL
Priority01000
NotifyWebClientUrlNULLPopulated

Note Priority is the cleanest single-column test — no nulls to reason about.

One poll cycle can fire many rules at once. Nine rule-driven requests were observed created within 80 milliseconds of each other.

Previous Article 102 - The Archive Database at a Glance
Next Article 201 - Two Ways an Archive Starts