[SERVER-30979] Run the fuzzer with CSRS primary stepdowns Created: 06/Sep/17  Updated: 30/Oct/23  Resolved: 01/Dec/17

Status: Closed
Project: Core Server
Component/s: Sharding, Testing Infrastructure
Affects Version/s: None
Fix Version/s: 3.6.4, 3.7.1

Type: Task Priority: Major - P3
Reporter: Max Hirschhorn Assignee: Max Hirschhorn
Resolution: Fixed Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Backports
Depends
depends on SERVER-31128 run_validate_collections hook doesn't... Closed
depends on SERVER-31461 Resmoke stepdown hook should deal wit... Closed
is depended on by SERVER-31452 Run fuzzer with shard stepdowns Closed
Duplicate
is duplicated by SERVER-31307 Run jstestfuzz_sharded_continuous_ste... Closed
is duplicated by SERVER-24327 Run stepdowns in parallel with other ... Closed
Related
related to SERVER-40063 jstestfuzz_sharded_continuous_stepdow... Closed
related to SERVER-31173 Change resmoke sharded cluster shutdo... Closed
related to SERVER-34352 Run dbhash check before collection va... Closed
Backwards Compatibility: Fully Compatible
Backport Requested:
v3.6
Sprint: TIG 2017-10-02, TIG 2017-10-23, TIG 2017-11-13, TIG 2017-12-04, TIG 2017-12-18
Participants:
Linked BF Score: 26

 Description   

A new resmoke.py YAML suite and Evergreen tasks should be defined to run the fuzzer with CSRS primary stepdowns happening in the background. The YAML suite should be adapted from the existing jstestfuzz_sharded.yml and causally_consistent_jscore_passthrough.yml suites where it'll run tests from the jstestfuzz/out/ directory against a sharded cluster with two replica set shards. Additionally, the new suite should define TestData.runningWithConfigStepdowns=true option and TestData.runningWithShardStepdowns=true so that the fuzzer can implement configuration-specific blacklisting similar to what is done with TestData.ignoreCommandsIncompatibleWithInitialSync=true in the jstestfuzz_replication_initsync.yml test suite.

The new suite should specify all_nodes_electable=true for the config server replica set started by resmoke.py (see SERVER-30683 and SERVER-30684) as well as enable the StepdownThread hook with config_stepdown=true and shard_stepdown=true (see SERVER-30685).

Note: It may be advantageous to initially commit a version of the jstestfuzz_sharded_continuous_stepdown.yml test suite that runs with config_stepdown=true and shard_stepdown=false (and the follow-up with another commit on this ticket for changing shard_stepdown to true) in case it is difficult to stabilize both simultaneously.

Names of new resmoke.py YAML suites
  • jstestfuzz_sharded_continuous_stepdown.yml
Names of new Evergreen tasks
  • jstestfuzz_sharded_continuous_stepdown
  • jstestfuzz_sharded_continuous_stepdown_WT
  • jstestfuzz_concurrent_sharded_continuous_stepdown
  • jstestfuzz_concurrent_sharded_continuous_stepdown_WT

The new Evergreen tasks should be run on the following build variants:

  • Enterprise RHEL 6.2
  • Enterprise RHEL 6.2 (inMemory)
  • SSL Windows 2008R2
  • SSL OS X 10.10
  • ASAN SSL Ubuntu 16.04


 Comments   
Comment by Githook User [ 10/Mar/18 ]

Author:

{'email': 'max.hirschhorn@mongodb.com', 'name': 'Max Hirschhorn', 'username': 'visemet'}

Message: SERVER-30979 Run the fuzzer with CSRS primary stepdowns.

(cherry picked from commit 26279e86956158edf69e8ed4e82cc211bf90ede4)
Branch: v3.6
https://github.com/mongodb/mongo/commit/2394d07abe45037f44e0cdff7a56abb92e86f0a6

Comment by Githook User [ 01/Dec/17 ]

Author:

{'name': 'Max Hirschhorn', 'username': 'visemet', 'email': 'max.hirschhorn@mongodb.com'}

Message: SERVER-30979 Run the fuzzer with CSRS primary stepdowns.
Branch: master
https://github.com/mongodb/mongo/commit/26279e86956158edf69e8ed4e82cc211bf90ede4

Generated at Thu Feb 08 04:25:37 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.