Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Allow task_table_update to fail when tasks are finished. #1927

Merged
merged 2 commits into from
Apr 20, 2018
Merged

Allow task_table_update to fail when tasks are finished. #1927

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
0c2a95a
Allow task_table_update to fail when tasks are finished.
robertnishihara Apr 20, 2018
e4a5328
Add comment.
robertnishihara Apr 20, 2018
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 4 additions & 6 deletions src/common/state/redis.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -966,18 +966,16 @@ void redis_task_table_update_callback(redisAsyncContext *c,
// task table update can race with the liveness monitor. Do not retry the
// update unless the caller is sure that the receiving subscriber is still
// alive in the db_client table.
if (reply->type == REDIS_REPLY_ERROR &&
strcmp(reply->str, "No subscribers received message.") == 0) {
RAY_LOG(WARNING) << "No subscribers received the task_table_update "
<< "message.";
if (reply->type == REDIS_REPLY_ERROR) {
RAY_LOG(WARNING) << "task_table_update failed with " << reply->str;
if (callback_data->retry.fail_callback != NULL) {
callback_data->retry.fail_callback(callback_data->id,
callback_data->user_context,
callback_data->data->Get());
} else {
RAY_LOG(FATAL) << "task_table_update failed and no fail_callback is set";
}
} else {
RAY_CHECK(reply->type != REDIS_REPLY_ERROR) << "reply->str is "
<< reply->str;
RAY_CHECK(strcmp(reply->str, "OK") == 0) << "reply->str is " << reply->str;

/* Call the done callback if there is one. */
Expand Down
16 changes: 15 additions & 1 deletion src/local_scheduler/local_scheduler.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -565,6 +565,11 @@ void assign_task_to_worker(LocalSchedulerState *state,
}
}

// This is used to allow task_table_update to fail.
void allow_task_table_update_failure(UniqueID id,
void *user_context,
void *user_data) {}

void finish_task(LocalSchedulerState *state, LocalSchedulerClient *worker) {
if (worker->task_in_progress != NULL) {
TaskSpec *spec = Task_task_execution_spec(worker->task_in_progress)->Spec();
Expand Down Expand Up @@ -622,7 +627,16 @@ void finish_task(LocalSchedulerState *state, LocalSchedulerClient *worker) {
int task_state = TASK_STATUS_DONE;
Task_set_state(worker->task_in_progress, task_state);
#if !RAY_USE_NEW_GCS
task_table_update(state->db, worker->task_in_progress, NULL, NULL, NULL);
auto retryInfo = RetryInfo{
.num_retries = 0, // This value is unused.
.timeout = 0, // This value is unused.
.fail_callback = allow_task_table_update_failure,
};

// We allow this call to fail in case the driver has been removed and the
// task table entries have already been cleaned up by the monitor.
task_table_update(state->db, worker->task_in_progress, &retryInfo, NULL,
NULL);
#else
RAY_CHECK_OK(TaskTableAdd(&state->gcs_client, worker->task_in_progress));
Task_free(worker->task_in_progress);
Expand Down