ruojieranyishen commented on code in PR #2195:
URL:
https://github.com/apache/incubator-pegasus/pull/2195#discussion_r1986734402
##########
src/meta/meta_bulk_load_service.cpp:
##########
@@ -587,11 +587,14 @@ void
bulk_load_service::try_resend_bulk_load_request(const std::string &app_name
FAIL_POINT_INJECT_F("meta_bulk_load_resend_request", [](std::string_view)
{});
zauto_read_lock l(_lock);
if (is_app_bulk_loading_unlocked(pid.get_app_id())) {
+ int32_t interval = _partition_bulk_load_info[pid].status ==
bulk_load_status::BLS_INGESTING
+ ?
bulk_load_constant::BULK_LOAD_INGEST_REQUEST_INTERVAL
+ :
bulk_load_constant::BULK_LOAD_REQUEST_INTERVAL;
tasking::enqueue(LPC_META_STATE_NORMAL,
_meta_svc->tracker(),
std::bind(&bulk_load_service::partition_bulk_load,
this, app_name, pid),
0,
-
std::chrono::seconds(bulk_load_constant::BULK_LOAD_REQUEST_INTERVAL));
+ std::chrono::milliseconds(interval));
Review Comment:
The partition_bulk_load function prompts the replica to check if the RocksDB
ingest is finished. Shortening its execution interval can reduce the
write-blocking time of the partition, from an average of 30 seconds to 500
milliseconds.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
