-
Type:
Task
-
Resolution: Done
-
Priority:
Minor - P4
-
None
-
Affects Version/s: None
-
Component/s: Storage
-
None
-
Storage Execution
-
None
-
None
-
None
-
None
-
None
-
None
-
None
1. Avoid letting the server continue running if RecoveryUnit::waitUntilDurable() return false.
if (MONGO_FAIL_POINT(crashAfterStartingIndexBuild)) {
log() << "Index build interrupted due to 'crashAfterStartingIndexBuild' failpoint. Exiting "
"after waiting for changes to become durable.";
Locker::LockSnapshot lockInfo;
_txn->lockState()->saveLockStateAndUnlock(&lockInfo);
if (_txn->recoveryUnit()->waitUntilDurable()) {
quickExit(EXIT_TEST);
}
}
2. Inline the contents of the abortDuringIndexBuild() function and eliminate the unnecessary usage of startParallelShell() in index_retry.js and index_no_retry.js. The changes from SERVER-23510 made it so that creating the background index should synchronously crash the server, so the tests no longer need to use multiple clients.
- is related to
-
SERVER-23510 Fix index_retry.js and index_no_retry.js to not be timing-based
-
- Closed
-