Commit 6e5975b8 authored by Zardosht Kasheff's avatar Zardosht Kasheff Committed by Yoni Fogel

[t:4876], some comments

git-svn-id: file:///svn/toku/tokudb@44340 c7de825b-a66e-492c-adef-691d508d4ae1
parent dc2d97cc
......@@ -166,7 +166,7 @@ struct tokutxn {
toku_mutex_t txn_lock;
// Protected by the txn lock:
OMT open_fts; // a collection of the brts that we touched. Indexed by filenum.
OMT open_fts; // a collection of the fts that we touched. Indexed by filenum.
struct txn_roll_info roll_info; // Info used to manage rollback entries
// Protected by the txn manager lock:
......
......@@ -321,10 +321,10 @@ int toku_txn_manager_start_txn(
bool for_recovery)
{
int r;
// we take the txn_manager_lock before writing to the log
// we may be able to move this lock acquisition
// down to just before inserting into logger->live_txns
// if we know the caller has the multi operation lock
// we take the txn_manager_lock before writing to the log,
// because the act of getting a transaction ID and adding the
// txn to the proper OMTs must be atomic. MVCC depends
// on this.
toku_mutex_lock(&txn_manager->txn_manager_lock);
if (garbage_collection_debug) {
verify_snapshot_system(txn_manager);
......
......@@ -107,41 +107,16 @@ toku_txn_commit_only(DB_TXN * txn, u_int32_t flags,
TOKULOGGER logger = txn->mgrp->i->logger;
LSN do_fsync_lsn;
BOOL do_fsync;
//
// quickie fix for 5.2.0, need to extract these variables so that
// we can do the fsync after the close of txn. We need to do it
// after the close because if we do it before, there are race
// conditions exposed by test_stress1.c (#4145, #4153) // release locks after completing the txn
//
// TODO: (Zardosht) refine this comment
//
// Here is what was going on. In Maxwell (5.1.X), we used to
// call toku_txn_maybe_fsync_log in between toku_txn_release_locks
// and toku_txn_close_txn. As a result, the ydb lock was released
// and retaken in between these two calls. This was wrong, as the
// two commands need to be atomic. The problem was that
// when the ydb lock was released, the locks that this txn took
// were released, but the txn was not removed from the list of
// live transactions. This allowed the following sequence of events:
// - another txn B comes and writes to some key this transaction wrote to
// - txn B successfully commits
// - read txn C comes along, sees this transaction in its live list,
// but NOT txn B, which came after this transaction.
// This is incorrect. When txn C comes across a leafentry that has been
// modified by both this transaction and B, it'll read B's value, even
// though it cannot read this transaction's value, which comes below
// B's value on the leafentry's stack. This behavior is incorrect.
// This causes a failure in the test_stress tests.
//
toku_txn_get_fsync_info(ttxn, &do_fsync, &do_fsync_lsn);
// remove the txn from the list of live transactions, and then
// release the lock tree locks. MVCC requires that toku_txn_complete_txn
// get called first, otherwise we have bugs, such as #4145 and #4153
toku_txn_complete_txn(ttxn);
r = toku_txn_release_locks(txn);
// this lock must be released after toku_txn_complete_txn because
// this lock must be held until the references to the open FTs is released
// begin checkpoint logs these associations, so we must be protect
// the changing of these associations with checkpointing
// Close the logger after releasing the locks
r = toku_txn_release_locks(txn);
if (release_multi_operation_client_lock) {
toku_multi_operation_client_unlock();
}
......@@ -278,8 +253,6 @@ locked_txn_id(DB_TXN *txn) {
static int
toku_txn_txn_stat (DB_TXN *txn, struct txn_stat **txn_stat) {
XMALLOC(*txn_stat);
// TODO: (Zardosht) make sure thread safety of this is resolved
// with some tokutxn lock that Leif is working on
return toku_logger_txn_rollback_raw_count(db_txn_struct_i(txn)->tokutxn, &(*txn_stat)->rollback_raw_count);
}
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment