EX294 Coordinate rolling updates Practice Test 3 — 15 Questions

Question 1

You are writing an Ansible playbook to perform a rolling update of a load-balanced web server fleet. You want to ensure that each host is removed from the load balancer before being updated and added back after the update, and that the update proceeds in small batches to minimize user impact. Which TWO play-level keywords or strategies should you use to achieve this? (Choose two.)

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

`pre_tasks` and `post_tasks` to handle load balancer removal and addition. To perform a controlled rolling update with load balancer integration, you use `serial` to limit the batch size and `pre_tasks`/`post_tasks` to remove and add hosts from the load balancer per batch. This ensures that only a small number of hosts are out of service at a time and that each host is properly drained before update and restored after.

Answer

`ignore_errors: true` to continue despite failures

Answer

`strategy: free` to allow hosts to proceed independently

Answer

`run_once: true` on the load balancer tasks

Question 2

A platform engineer runs an Ansible playbook against a 20-node web fleet to install a new kernel. The playbook defines serial: "25%" and max_fail_percentage: 10. During the first batch, one host fails its package transaction and becomes unreachable. What happens to the playbook execution?

Accepted Answer

The playbook completes the current batch and then stops before starting the next batch.. With 20 hosts and serial: "25%", Ansible processes 5 hosts per batch. max_fail_percentage: 10 is applied per batch, so the batch can tolerate zero failures because 10% of 5 rounds down to 0. One failed host is 20% of the batch, which exceeds the limit, causing the play to stop after the current batch completes. This protects the fleet from a widespread bad update.

Answer

The playbook ignores the failure and continues to the next batch because max_fail_percentage is only checked at the end of the play.

Answer

The playbook aborts immediately because any host failure exceeds max_fail_percentage.

Answer

The playbook retries the failed host automatically before evaluating max_fail_percentage.

Question 3

You are designing a rolling update playbook for a 20-node web cluster. The application requires that at least 15 nodes remain online at all times. You plan to use `serial` to control batch size. Which TWO of the following `serial` values would violate the requirement if a batch fails and Ansible stops processing further batches? (Choose two.)

Accepted Answer

serial: 7. The requirement is that at least 15 of 20 nodes must remain online. If a batch fails, the nodes in that batch may become unavailable. Therefore, the batch size must not exceed 5, because 20 - 5 = 15. A `serial` value of 6 or 7 would leave 14 or 13 nodes online respectively, violating the requirement. Values of 3, 4, or 5 are safe. The correct answers are the values that exceed the safe threshold.

Answer

serial: 4

Answer

serial: 5

Answer

serial: 3

Question 4

You are designing an Ansible playbook for a rolling update of a 10-node RHEL application cluster. The application can tolerate up to three nodes being offline simultaneously, and you want the playbook to stop the entire rollout if more than 20 percent of the hosts in the current batch fail. Which two play configurations satisfy these requirements? (Choose two.)

Accepted Answer

`serial: 3` with `max_fail_percentage: 20`. The availability constraint allows at most three nodes offline, so the serial batch size must be three hosts, expressed either as an integer or as 30 percent of a 10-node inventory. The failure policy requires aborting when more than 20 percent of the current batch fails, which for a three-host batch means any failure triggers an abort. Both `serial: 3` and `serial: 30%` combined with `max_fail_percentage: 20` meet these conditions.

Answer

`serial: 4` with `max_fail_percentage: 20`

Answer

`serial: 2` with `max_fail_percentage: 20`

Answer

`serial: 3` with `max_fail_percentage: 30`

Question 5

An administrator is using an Ansible playbook to perform a rolling update on a group of web servers. The playbook uses `serial: 2`. After the first batch of two servers is updated, the administrator notices that the playbook immediately proceeds to update the next batch without waiting for the first batch to complete their post-update tasks. Which Ansible behavior explains this?

Accepted Answer

The playbook uses the default `linear` strategy, which waits for all hosts in the current batch to complete before moving to the next batch.. In Ansible, the default strategy is linear, which ensures that all hosts in the current batch complete each task before the next task begins. When serial is used, the playbook processes the first batch entirely, including post-update tasks, before starting the next batch. This behavior guarantees that each batch is fully updated and any post-update steps are completed before moving on, which is critical for rolling updates.

Answer

The playbook has `async` enabled on the post-update tasks, causing them to run in the background.

Answer

The playbook uses `serial: 2` with a `max_fail_percentage` that is too high, causing it to skip waiting.

Answer

The playbook is configured with `strategy: free`, allowing hosts to proceed independently.

Question 6

You are writing an Ansible playbook to perform a rolling update of a load-balanced web server fleet. You want to ensure that each host is removed from the load balancer before being updated and added back after the update, and that the update proceeds in small batches to minimize user impact. Which TWO play-level keywords or strategies should you use to achieve this? (Choose two.)

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

serial: 3 with a `pause` module in a separate play that targets `localhost`.. The `pause` module can be used in the `post_tasks` section of a play that uses `serial`. The `post_tasks` run after each batch of hosts, so placing a `pause` task there will introduce a delay between batches. This is the standard way to pause between serial batches. Using a separate play targeting localhost would only pause once, not between batches.

Answer

serial: 3 with a `pause` module at the end of the play.

Answer

serial: 3 with a `delay` keyword on the update task.

Answer

serial: 3 with a `wait_for` task after the update tasks.

Question 7

You are performing a rolling update of a 6-node cluster using Ansible with `serial: 2`. The playbook includes a task that removes a node from a load balancer, updates the application, and then adds the node back. You want to ensure that if the load balancer removal task fails on any host, the playbook stops immediately for that host and does not attempt the update, but other hosts in the same batch should continue. Which task-level keyword should you use on the load balancer removal task?

Accepted Answer

block with rescue. To stop processing for a specific host on failure while allowing other hosts to continue, you can use a `block` with a `rescue` that calls `meta: end_host`. This effectively removes the host from the play without affecting other hosts. The other options either ignore the failure and proceed dangerously, or affect all hosts globally. The `block`/`rescue` structure provides per-host error handling.

Answer

failed_when: false

Answer

any_errors_fatal: true

Answer

ignore_errors: yes

Question 8

You are updating a fleet of 12 Red Hat Enterprise Linux 9 web servers with an Ansible playbook that uses serial: 3. During the second batch, a task fails on one host because a required package is not available in the configured repository. The playbook aborts immediately, leaving the remaining batches unmanaged. You want the playbook to continue processing the remaining batches while excluding that failed host, and to be able to re-run the playbook later only on hosts that still need the update. Which combination of settings should you add to the playbook?

Accepted Answer

Set max_fail_percentage to a value that tolerates the failure and use a retry file with --limit @retry_file on the next run.. Allowing a play to proceed past a batch failure while preserving a record of which hosts failed is the standard rolling-update pattern. max_fail_percentage sets the tolerance for failures within a serial batch, and the generated .retry file captures the failed hosts so a follow-up run with --limit @file targets only those systems instead of the whole fleet.

Answer

Set max_fail_percentage: 25 and use a separate inventory file that omits the failed host.

Answer

Add ignore_errors: true to every task and re-run the playbook against all hosts.

Answer

Use serial: 3 with a block/rescue that runs a task to remove the failed host from the in-memory inventory, then continue.

Question 9

An operations team uses an Ansible playbook with serial: 3 to update a five-node stateful cluster. They need the play to stop the entire rolling update if more than two hosts fail across all batches, not just within a single batch. Which combination of keywords achieves this?

Accepted Answer

Set serial: 3 with a block that uses a rescue section to track failures and fail the play when the count exceeds two.. max_fail_percentage and any_errors_fatal operate per batch or per task, not cumulatively across the whole rolling update. Because the requirement is a global limit of two failures across all batches, the playbook must implement its own failure tracking, such as a rescue section that increments a counter and fails the play when the count exceeds two. Simple keyword combinations cannot express that global threshold.

Answer

Set any_errors_fatal: true and serial: 3.

Answer

Set max_fail_percentage: 40 and serial: 5.

Answer

Set max_fail_percentage: 40 and serial: 3.

Question 10

A playbook performs a rolling update of a 20-node application cluster. The application can tolerate up to 4 nodes being offline at once. The playbook currently uses serial: 1, which is slow. You change it to serial: 5, but during testing you observe that 5 nodes go offline simultaneously. Which serial value should you use to meet the availability requirement while maximizing batch size?

Accepted Answer

serial: 4. The serial keyword defines how many hosts are processed per batch, and those hosts are the ones that may be offline during the update. To honor a four-node offline limit while updating as many hosts per batch as possible, the batch size must be exactly four, which is achieved with serial: 4.

Answer

serial: 10%

Answer

serial: 5

Answer

serial: 3

Question 11

A site reliability engineer is performing a rolling update on a 6-node cluster using serial: 2. The playbook must update the first two nodes, then pause for manual verification before continuing with the remaining nodes. Which approach correctly inserts the pause between batches?

Accepted Answer

Add a pause task with the prompt module at the end of the play, relying on serial to repeat it after each batch.. With serial: 2, Ansible repeats the whole play for each pair of hosts. A prompt or pause task placed at the end of the play therefore runs after every batch, creating a natural manual verification gate. Modules like wait_for and keywords like max_fail_percentage do not provide an operator-driven pause, so they cannot satisfy the requirement for manual approval between batches.

Answer

Use the wait_for module with a timeout between the two batches.

Answer

Set max_fail_percentage: 0 and rely on the failure to pause the playbook.

Answer

Set serial: 2 and add a when: inventory_hostname == groups['all'][1] condition to a pause task.

Question 12

An engineer is designing a rolling update playbook for a 9-node service cluster. The play must update three nodes per batch, continue even if up to one node in a batch fails, and must never run the update on more than three hosts concurrently. Which TWO directives should be included in the play? (Choose two.)

Accepted Answer

max_fail_percentage: 33. The play needs serial: 3 to create batches of three hosts and a failure threshold that permits one failure per batch. With three hosts per batch, a percentage of 34 or higher yields floor(3 * 0.34) = 1 allowed failure. any_errors_fatal is too strict, and throttle controls task workers rather than host batching. The two required directives are serial: 3 and a max_fail_percentage value above 33.

Answer

max_fail_percentage: 34

Answer

throttle: 3

Answer

any_errors_fatal: true

Question 13

A playbook updates a 9-node cluster using serial: 3. The first batch completes successfully. During the second batch, the play fails on all three hosts because a configuration file has a syntax error. You fix the template on the control node and want to resume the update without re-running the first batch and without touching the third batch until the second batch succeeds. What is the most efficient way to proceed?

Accepted Answer

Re-run the playbook with --limit set to the three hosts in the second batch, then run it again for the third batch.. The most efficient recovery is to limit the play to the hosts that failed, which avoids re-processing the successful first batch and defers the third batch until the second is verified. --limit accepts a host pattern or a retry file and is the intended mechanism for narrowing a run to specific systems.

Answer

Use the --start-at-task option to begin at the task that failed and re-run against the full inventory.

Answer

Re-run the playbook against the entire inventory with the same serial: 3 setting.

Answer

Change serial to 9 and re-run the playbook so all remaining hosts are updated together.

Question 14

An operations team is designing a rolling update for a stateful application that requires quorum (minimum 3 out of 5 nodes online). They plan to use Ansible's serial keyword. Which serial value ensures the update proceeds without breaking quorum while still being efficient?

Accepted Answer

serial: 2. Setting serial: 2 ensures that only 2 nodes are taken down at a time during the rolling update. With a quorum requirement of 3 out of 5 nodes, taking down 2 nodes leaves 3 online, maintaining quorum. This is the most efficient value that does not risk breaking quorum.

Answer

serial: 1

Answer

serial: 3

Answer

serial: 5

Question 15

Which TWO options are best practices for coordinating rolling updates with Ansible? (Choose exactly two.)

Accepted Answer

Use the serial keyword to update hosts in batches.. Option B is correct because the serial keyword in Ansible controls how many hosts are targeted per play iteration, allowing rolling updates in controlled batches (e.g., serial: 2 or serial: 25%) so that only a subset of hosts is updated at a time while the rest continue serving traffic. Option D is correct because max_fail_percentage defines a failure threshold within a serial batch; if the percentage of failed hosts exceeds that value, Ansible aborts the play, preventing a bad update from cascading across the entire fleet. Option A is not a best practice for rolling updates because ignore_errors: yes masks failures and lets the playbook proceed even when hosts are broken, defeating the safety purpose of batching. Option C is wrong because the default serial value is effectively all hosts in the play, which performs a simultaneous update rather than a rolling one. Option E is also wrong because running all hosts in parallel maximizes blast radius and downtime risk, the opposite of a rolling-update strategy.

Answer

Set ignore_errors: yes to ensure the playbook continues even if some hosts fail.

Answer

Use the default serial setting (all hosts) for simplicity.

Answer

Run all hosts in parallel to minimize total update time.