andreacavalli/rocksdb
0
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

WritePrepared: fix snapshot sequence in rollback (#4851)

Browse Source 
Summary:
The rollback algorithm in WritePrepared transactions requires reading the values before the transaction start. Currently it uses the prepare_seq -1 as the snapshot sequence number for the read. This is not correct since the passed sequence number must be for a valid snapshot. The patch fixes it by passing kMaxSequenceNumber instead. This is fine since all the writes done by the aborted transaction will be skipped during the read anyway.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/4851

Differential Revision: D13592773

Pulled By: maysamyabandeh

fbshipit-source-id: ff1bf92ea9909d4cccb173bdff49febc0e9eb7a2
This commit is contained in:
Maysam Yabandeh 2019-01-07 14:53:26 -08:00 committed by Facebook Github Bot
parent 42135523a0
commit 0ed98bf89e
1 changed files with 2 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
utilities/transactions/write_prepared_txn.cc
View File

@ -218,8 +218,7 @@ Status WritePreparedTxn::RollbackInternal() {
assert(GetId() > 0);
auto cf_map_shared_ptr = wpt_db_->GetCFHandleMap();
auto cf_comp_map_shared_ptr = wpt_db_->GetCFComparatorMap();
// In WritePrepared, the txn is is the same as prepare seq
auto last_visible_txn = GetId() - 1;
auto read_at_seq = kMaxSequenceNumber;
struct RollbackWriteBatchBuilder : public WriteBatch::Handler {
DBImpl* db_;
ReadOptions roptions;
@ -308,7 +307,7 @@ Status WritePreparedTxn::RollbackInternal() {
protected:
virtual bool WriteAfterCommit() const override { return false; }
} rollback_handler(db_impl_, wpt_db_, last_visible_txn, &rollback_batch,
} rollback_handler(db_impl_, wpt_db_, read_at_seq, &rollback_batch,
*cf_comp_map_shared_ptr.get(), *cf_map_shared_ptr.get(),
wpt_db_->txn_db_options_.rollback_merge_operands);
auto s = GetWriteBatch()->GetWriteBatch()->Iterate(&rollback_handler);