ClickHouse/src/Storages/MergeTree/MutatePlainMergeTreeTask.cpp

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

126 lines
3.7 KiB
C++
Raw Normal View History

#include <Storages/MergeTree/MutatePlainMergeTreeTask.h>
#include <Storages/StorageMergeTree.h>
2021-12-14 20:06:34 +00:00
#include <Interpreters/TransactionLog.h>
namespace DB
{
namespace ErrorCodes
{
extern const int LOGICAL_ERROR;
}
StorageID MutatePlainMergeTreeTask::getStorageID()
{
return storage.getStorageID();
}
void MutatePlainMergeTreeTask::onCompleted()
{
bool delay = state == State::SUCCESS;
task_result_callback(delay);
}
void MutatePlainMergeTreeTask::prepare()
{
future_part = merge_mutate_entry->future_part;
const Settings & settings = storage.getContext()->getSettingsRef();
merge_list_entry = storage.getContext()->getMergeList().insert(
storage.getStorageID(),
future_part,
Fix possible memory_tracker use-after-free for merges/mutations There are two possible cases for execution merges/mutations: 1) from background thread 2) from OPTIMIZE TABLE query 1) is pretty simple, it's memory tracking structure is as follow: current_thread::memory_tracker = level=Thread / description="(for thread)" == background_thread_memory_tracker = level=Thread / description="(for thread)" current_thread::memory_tracker.parent = level=Global / description="(total)" So as you can see it is pretty simple and MemoryTrackerThreadSwitcher does not do anything icky for this case. 2) is complex, it's memory tracking structure is as follow: current_thread::memory_tracker = level=Thread / description="(for thread)" current_thread::memory_tracker.parent = level=Process / description="(for query)" == background_thread_memory_tracker = level=Process / description="(for query)" Before this patch to track memory (and related things, like sampling, profiling and so on) for OPTIMIZE TABLE query dirty hacks was done to do this, since current_thread memory_tracker was of Thread scope, that does not have any limits. And so if will change parent for it to Merge/Mutate memory tracker (which also does not have some of settings) it will not be correctly tracked. To address this Merge/Mutate was set as parent not to the current_thread memory_tracker but to it's parent, since it's scope is Process with all settings. But that parent's memory_tracker is the memory_tracker of the thread_group, and so if you will have nested ThreadPool inside merge/mutate (this is the case for s3 async writes, which has been added in #33291) you may get use-after-free of memory_tracker. Consider the following example: MemoryTrackerThreadSwitcher() thread_group.memory_tracker.parent = merge_list_entry->memory_tracker (see also background_thread_memory_tracker above) CurrentThread::attachTo() current_thread.memory_tracker.parent = thread_group.memory_tracker CurrentThread::detachQuery() current_thread.memory_tracker.parent = thread_group.memory_tracker.parent # and this is equal to merge_list_entry->memory_tracker ~MemoryTrackerThreadSwitcher() thread_group.memory_tracker = thread_group.memory_tracker.parent So after the following we will get incorrect memory_tracker (from the mege_list_entry) when the next job in that ThreadPool will not have thread_group, since in this case it will not try to update the current_thread.memory_tracker.parent and use-after-free will happens. So to address the (2) issue, settings from the parent memory_tracker should be copied to the merge_list_entry->memory_tracker, to avoid playing with parent memory tracker. Note, that settings from the query (OPTIMIZE TABLE) is not available at that time, so it cannot be used (instead of parent's memory tracker settings). v2: remove memory_tracker.setOrRaiseHardLimit() from settings Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
2022-02-18 07:45:29 +00:00
settings);
stopwatch = std::make_unique<Stopwatch>();
write_part_log = [this] (const ExecutionStatus & execution_status)
{
mutate_task.reset();
storage.writePartLog(
PartLogElement::MUTATE_PART,
execution_status,
stopwatch->elapsed(),
future_part->name,
new_part,
future_part->parts,
merge_list_entry.get());
};
2021-10-04 21:13:18 +00:00
fake_query_context = Context::createCopy(storage.getContext());
fake_query_context->makeQueryContext();
fake_query_context->setCurrentQueryId("");
mutate_task = storage.merger_mutator.mutatePartToTemporaryPart(
future_part, metadata_snapshot, merge_mutate_entry->commands, merge_list_entry.get(),
2022-02-14 19:50:08 +00:00
time(nullptr), fake_query_context, merge_mutate_entry->txn, merge_mutate_entry->tagger->reserved_space, table_lock_holder);
}
bool MutatePlainMergeTreeTask::executeStep()
{
/// Make out memory tracker a parent of current thread memory tracker
MemoryTrackerThreadSwitcherPtr switcher;
if (merge_list_entry)
switcher = std::make_unique<MemoryTrackerThreadSwitcher>(*merge_list_entry);
switch (state)
{
2022-06-28 10:51:49 +00:00
case State::NEED_PREPARE:
{
prepare();
state = State::NEED_EXECUTE;
return true;
}
2022-06-28 10:51:49 +00:00
case State::NEED_EXECUTE:
{
try
{
if (mutate_task->execute())
return true;
new_part = mutate_task->getFuture().get();
2022-06-24 11:19:29 +00:00
MergeTreeData::Transaction transaction(storage, merge_mutate_entry->txn.get());
2022-03-09 20:38:18 +00:00
/// FIXME Transactions: it's too optimistic, better to lock parts before starting transaction
2022-10-22 22:51:59 +00:00
storage.renameTempPartAndReplace(new_part, transaction);
2022-06-24 11:19:29 +00:00
transaction.commit();
storage.updateMutationEntriesErrors(future_part, true, "");
write_part_log({});
state = State::NEED_FINISH;
return true;
}
catch (...)
{
2021-12-14 20:06:34 +00:00
if (merge_mutate_entry->txn)
merge_mutate_entry->txn->onException();
2022-01-19 18:29:31 +00:00
String exception_message = getCurrentExceptionMessage(false);
2022-02-03 18:57:09 +00:00
LOG_ERROR(&Poco::Logger::get("MutatePlainMergeTreeTask"), "{}", exception_message);
2022-01-19 18:29:31 +00:00
storage.updateMutationEntriesErrors(future_part, false, exception_message);
write_part_log(ExecutionStatus::fromCurrentException());
tryLogCurrentException(__PRETTY_FUNCTION__);
return false;
}
}
case State::NEED_FINISH :
{
// Nothing to do
state = State::SUCCESS;
return false;
}
case State::SUCCESS:
{
throw Exception(ErrorCodes::LOGICAL_ERROR, "Task with state SUCCESS mustn't be executed again");
}
}
return false;
}
}