-
Type:
Bug
-
Resolution: Unresolved
-
Priority:
Major - P3
-
None
-
Affects Version/s: None
-
Component/s: None
-
Storage Engines
-
StorEng - Defined Pipeline
A customer system has seen repeated crashes with this stack trace:
1 mongo::printStackTrace() 2 abruptQuitWithAddrSignal 3 __kernel_rt_sigreturn 4 __wt_conf_compile_api_call 5 __session_begin_transaction 6 mongo::WiredTigerBeginTxnBlock::WiredTigerBeginTxnBlock(mongo::WiredTigerSession*, mongo::PrepareConflictBehavior, mongo::WiredTigerBeginTxnBlock::RoundUpPreparedTimestamps, mongo::WiredTigerBeginTxnBlock::RoundUpReadTimestamp, mongo::RecoveryUnit::UntimestampedWriteAssertionLevel) 7 mongo::WiredTigerRecoveryUnit::_txnOpen() 8 mongo::WiredTigerRecoveryUnit::preallocateSnapshot() 9 mongo::shard_role_details::SnapshotAttempt::openStorageSnapshot()
It looks like the new(ish) config compilation code is generating the fault when called in WT_SESSION::begin_transaction().
I can't tell from the context whether the configuration we are getting here is valid, but either way we should be able to handle it without failing.