Tableau SCA-C01 - Tableau Server Certified Associate Exam

In an HA Tableau Server cluster, theCoordination Service(ZooKeeper ensemble) maintains cluster state—let’s find the best way to check it:

Coordination Service:

Runs on multiple nodes (3 or 5 in HA) to ensure quorum and failover.

Status indicates if it’s running and synced—critical for cluster health.

Option C (Run tsm status -v): Correct.

Details: tsm status --verbose lists all processes across nodes, including Coordination Service (e.g., "Coordination Service: RUNNING").

Why Best: Provides detailed, node-specific status in the CLI—e.g., "Node 1: RUNNING, Node 2: RUNNING."

Use: Run on the initial node; -v ensures full output.

Option A (TSM web client Status page): Incorrect.

Why: The TSM UI (Server > Status) shows process counts (e.g., "Coordination Service: 3 instances"), but not detailed per-node status—less granular than CLI.

Option B (tsm maintenance ziplogs): Incorrect.

Why: Generates log archives for troubleshooting, not a real-time status check.

Option D (Tableau Server Status page): Incorrect.

Why: The Server Status page (Server > Statusin the web UI) monitors application processes (e.g., VizQL), not TSM’s Coordination Service.

Why This Matters: Coordination Service health ensures HA stability—tsm status -v is the admin’s go-to for precision.

Tableau Server’s performance and stability depend on dedicated resources and proper configuration. Running additional software on the same server is the most likely to cause problems because:

Resource contention: Tableau Server requires significant CPU, RAM, and disk I/O. Other software (e.g., databases, web servers) can compete for these resources, leading to slowdowns, crashes, or failed tasks.

Port conflicts: Tableau uses specific ports (e.g., 80, 443, 8850), and other applications might interfere.

Security risks: Additional software increases the attack surface, potentially compromising Tableau Server.

Tableau recommends running the server on dedicated hardware without unrelated applications.

Option A (Running additional software on the server): Correct. This is a common cause of performance issues and is explicitly discouraged in Tableau’s best practices.

Option B (Separating the Backgrounder and VizQL processes to different machines): Incorrect. This is a supported multi-node configuration that can improve performance, not cause problems, if properly set up via TSM.

Option C (Configuring the server to use a static IP address): Incorrect. A static IP is recommended for Tableau Server to ensure consistent network access, so it’s unlikely to cause issues.

Option D (Using a non-default installation path): Incorrect. While not default, a custom path is supported (via TSM or installer options) and unlikely to cause problems if permissions and disk space are adequate.

Abandoned workbooks—those no longer actively used—can clutter Tableau Server and consume resources (e.g., disk space for extracts). Identifying them efficiently requires leveraging built-in administrative tools rather than manual or destructive methods. Let’s explore this in depth:

Tableau Server Admin Views: Tableau provides pre-built administrative views to monitor server health, usage, and content. TheStale Contentview, accessible underServer > Status > Administrative Views, is specifically designed to identify content (workbooks, data sources) that hasn’t been viewed or modified recently. It shows:

Content name, owner, and project.

Last viewed date and last modified date.

View count over a period.This view uses Repository data to track usage metrics, making it ideal for spotting abandoned workbooks.

Option A (Use the Stale Content administrative view): Correct. This is the most efficient and non-invasive method. You can filter by last viewed date (e.g., >30 days ago) to identify workbooks with large extracts that users have stopped accessing. From there, you can contact owners or delete the content if policy allows. It’s a server administrator’s go-to tool for content management.

Option B (Examine extract files in ProgramData/.../extract): Incorrect and impractical. The ProgramData/Tableau/Tableau Server/data/tabsvc/dataengine/extract directory stores .hyper extract files, but:

File names are cryptic (e.g., GUIDs), not tied directly to workbook names.

It doesn’t indicate usage or abandonment—only file presence and size.

Manual inspection is time-consuming and error-prone compared to the Stale Content view.

Option C (Delete all extracts and allow them to be re-generated): Incorrect and risky. Deleting extracts (e.g., via tsm maintenance cleanup) removes them without identifying usage. Regeneration only occurs if a schedule or user triggers it, potentially disrupting active users and losing historical data unnecessarily.

Option D (View all workbooks, and sort by the Modified date): Partially effective but inefficient. In the Tableau Server UI (e.g., underContent > Workbooks), you can sort by "Last Modified," but:

It doesn’t show viewership (a workbook might be modified recently but unused).

It’s manual and doesn’t scale for large deployments compared to the Stale Content view.

Why This Matters: The Stale Content view leverages Tableau’s metadata to provide actionable insights, saving time and reducing risk compared to manual or destructive alternatives. It’s part of Tableau’s governance toolkit.

Tableau Server relies on several processes to function, each with a specific role. TheRepositoryprocess (powered by PostgreSQL) serves as the database for metadata, storing critical information such as:

User and group details.

Permissions and site configurations.

Workbook and data source metadata (e.g., schedules, subscriptions).

Option B (Repository): Correct. The Repository is the centralized database that holds all metadata, making it the backbone of Tableau Server’s content management. There are typically two instances in an HA setup (one active, one passive), monitored by the Cluster Controller.

Option A (Data Engine): Incorrect. The Data Engine manages in-memory data processing and extract storage (e.g., .hyper files), not metadata. It’s separate from the Repository.

Option C (Backgrounder): Incorrect. The Backgrounder handles background tasks like extract refreshes and subscriptions, but it doesn’t store metadata—it interacts with the Repository to retrieve task details.

Option D (File Store): Incorrect. The File Store manages physical extract files and workbook assets, not metadata, which is stored in the Repository.

In a high-availability (HA) Tableau Server setup, theRepository(PostgreSQL) has an active and passive instance. Failover occurs if the active Repository fails. Let’s dive into the process:

HA Setup:

Two Repository instances across nodes (active/passive).

Failover switches to the passive instance if the active one becomes unavailable (e.g., crash, network issue).

Cluster Controller:

Role: Monitors all processes (e.g., Repository, File Store) across nodes, detecting failures via heartbeats and status checks.

Failover Decision: If the active Repository stops responding, Cluster Controller initiates failover, promoting the passive instance to active.

Coordination: Works with Coordination Service (ZooKeeper) to update topology but makes the initial detection call.

Option A (Cluster Controller): Correct.

Why: It’s the watchdog process, constantly monitoring Repository health and triggering failover when needed.

Option B (Coordination Service): Incorrect.

Role: ZooKeeper maintains cluster state and coordinates topology updates post-failover, but doesn’t detect the failure—Cluster Controller does.

Option C (Gateway): Incorrect.

Role: Routes client requests—unrelated to internal process monitoring or failover.

Option D (Backgrounder): Incorrect.

Role: Executes background tasks—no involvement in Repository failover decisions.

Why This Matters: Understanding failover ensures HA reliability—Cluster Controller is the linchpin for resilience.

Tableau Server’s search functionality relies on an indexed catalog of content (workbooks, data sources, etc.) stored in the Repository. If a user can see a workbook in the UI (e.g., underContent > Workbooks) but not find it via search, the search index may be outdated or corrupted. This can happen due to:

Indexing delays after publishing.

Server maintenance or crashes affecting the index.

Option D (Run the tsm maintenance reindex-search command): Correct. This command rebuilds the search index, ensuring all content (including the user’s workbook) is properly cataloged and searchable. Steps:

Stop Tableau Server (tsm stop).

Run tsm maintenance reindex-search.

Start Tableau Server (tsm start).This is a server administrator task and resolves systemic search issues.

Option A (Re-publish the workbook with keywords): Incorrect. Re-publishing might update the index for that workbook, but it doesn’t fix a broader indexing problem. Keywords enhance relevance, not indexing itself.

Option B (Add tags to the workbook): Incorrect. Tags improve searchability but don’t address an index failure. If the workbook isn’t indexed, tags won’t help.

Option C (Log out, and then log back in): Incorrect. This refreshes the user session but doesn’t affect the server-side search index.

Why This Matters: A reliable search index is critical for content discovery in large deployments—reindex-search ensures consistency.

Tableau Server’s installer configures process instances based on hardware and deployment type (single-node vs. multi-node). For a single-node installation with 4 cores, we need to consider thedefault process topology. Let’s break this down exhaustively:

Key Processes:

Gateway: Handles incoming requests (1 instance).

Application Server (VizPortal): Manages UI and sessions (1 instance).

VizQL Server: Renders visualizations (2 instances).

Backgrounder: Runs extract refreshes, subscriptions (1 instance).

Data Server: Manages data connections (1 instance).

File Store: Stores extracts (1 instance).

Repository: Metadata database (1 instance, active).

Cluster Controller, Cache Server, etc.: Supporting processes (typically 1 each).

Default Configuration:

On a single-node install, Tableau sets1 instance per processunless specified otherwise, except for VizQL, which defaults to 2.

The installer doesn’t scale instances linearly with cores (e.g., 4 cores ≠ 4 instances). Post-install, TSM can adjust this (e.g., tsm topology set-process), but the question asks for theinstalled default.

Minimum hardware (8 cores, 32 GB RAM) suggests higher defaults, but 4 cores still triggers a minimal setup.

Option B (1): Correct with Caveat.

Most processes (e.g., Backgrounder, Gateway, Data Server) default to 1 instance on install, regardless of 4 cores.

VizQL defaults to 2, but the question’s phrasing ("each process") implies a general rule. Historically (and per docs), 1 is the baseline for most, with VizQL as the exception.

Interpretation: Assuming "each" means the typical case, 1 fits most processes on a 4-core single-node setup.

Option A (2): Incorrect. Only VizQL defaults to 2; others don’t.

Option C (8): Incorrect. Far exceeds defaults—8 cores might justify more, but not 4.

Option D (4): Incorrect. Not tied to core count by default; manual config would be needed.

Why This Matters: Understanding defaults aids capacity planning—4 cores is below production minimum (8), so performance tuning may be needed post-install.

Tableau Server supports multi-tenancy via sites, each with customizable settings managed by server or site administrators. Let’s analyze what’s configurable per site:

Site Settings: Found in the web UI underSite > Settings > General. Server admins can override site admin settings.

Option B (Limitation on storage space): Correct.

Details: Server admins can set astorage quotaper site (e.g., 100 GB) to cap disk usage for extracts and workbooks.

How: In TSM or site settings (if enabled)—e.g., tsm configuration set -k site.storage.quota -v 100000.

Impact: Prevents one site from monopolizing resources in multi-site deployments.

Option D (Language and locale): Correct.

Details: Each site can set itslanguage(e.g., English, French) andlocale(e.g., date/number formats).

How: Site settings UI—e.g., "Language: French, Locale: France."

Impact: Tailors the user experience per site’s audience.

Option A (Ability to embed credentials): Incorrect.

Details: Embedding credentials (e.g., in data sources) is a server-wide setting (tsm data-access), not per-site. Site admins can’t override it.

Option C (Limitation on number of users): Incorrect.

Details: User limits are tied to licenses (server-wide), not configurable per site. Site admins manage user assignments, not quotas.

Why This Matters: Site-specific settings enable tailored governance and resource allocation in multi-tenant environments.

When a user reports slow performance for a newly-published workbook on Tableau Server, troubleshooting requires isolating the cause—e.g., data source issues, server load, workbook design, or caching. Thefirst questionshould establish a baseline to narrow the scope. Let’s analyze this step-by-step with depth:

Performance Context:

A workbook’s speed depends on:

Data Source: Query complexity, size, network latency (e.g., database vs. extract).

Workbook Design: Filters, calculations, dashboard complexity.

Server Resources: VizQL rendering, Backgrounder load, caching.

"Newly-published" implies it’s not yet optimized or cached on the server.

Option A (Does it run any faster in Tableau Desktop?): Correct.

Why First: Comparing Desktop vs. Server performance is the most foundational diagnostic step:

Desktop Baseline: If it’s slow in Desktop (local machine), the issue likely lies in the workbook (e.g., complex queries, large data) or data source (e.g., slow database)—not Server-specific.

Server Difference: If it’s fast in Desktop but slow on Server, the problem could be server-side (e.g., resource contention, network latency to the data source from Server).

Practical Next Steps:

Slow in Desktop: Optimize workbook (e.g., simplify calcs, use extracts).

Fast in Desktop: Check Server (e.g., caching, VizQL load).

Why Critical: Establishes whether the issue is inherent to the workbook/data or introduced by Server—guides all further investigation.

Option B (Does the workbook always run slowly or does performance vary?): Useful but secondary.

Why Not First: Variability (e.g., slow at peak times) points to server load, but without a Desktop baseline, you can’t rule out workbook design. It’s a follow-up question after A.

Detail: Variability might suggest caching or concurrent user impact, but it assumes Server-side causation prematurely.

Option C (How many times have you opened the workbook in Tableau Server?): Less relevant initially.

Why Not First: Frequency of access might affect caching (first load is slower, subsequent loads faster), but it’s too specific and doesn’t isolate Desktop vs. Server. It’s a niche follow-up.

Option D (Did you enable caching on the workbook?): Misleading and incorrect.

Why Not First: Caching is server-managed (e.g., VizQL cache settings via tsm data-access caching set), not a user-toggle per workbook. Users don’t "enable" it—admins do. Plus, it’s premature without a baseline.

Why This Matters: Starting with Desktop performance cuts through assumptions, pinpointing whether the root cause is client-side (workbook/data) or server-side—essential for efficient resolution in production.