hcq/oceanbase
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

[CP] fix truncate timeout may be larger than 1000s

Browse Source
This commit is contained in:
obdev
2023-08-01 18:18:22 +00:00
committed by ob-robot
parent 4238db272a
commit 6e155c8467
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/sql/engine/cmd/ob_table_executor.cpp
View File

@ -2146,13 +2146,13 @@ int ObTruncateTableExecutor::execute(ObExecContext &ctx, ObTruncateTableStmt &st
} else {
// new parallel truncate
ObTimeoutCtx ctx;
if (OB_FAIL(ObShareUtil::set_default_timeout_ctx(ctx, GCONF._ob_ddl_timeout))) {
if (OB_FAIL(ObShareUtil::set_default_timeout_ctx(ctx, (static_cast<obrpc::ObRpcProxy*>(common_rpc_proxy))->timeout()))) {
LOG_WARN("fail to set timeout ctx", K(ret));
} else {
int64_t start_time = ObTimeUtility::current_time();
while (OB_SUCC(ret)) {
DEBUG_SYNC(BEFORE_PARELLEL_TRUNCATE);
if (OB_FAIL(common_rpc_proxy->truncate_table_v2(truncate_table_arg, res))) {
if (OB_FAIL(common_rpc_proxy->timeout(ctx.get_timeout()).truncate_table_v2(truncate_table_arg, res))) {
LOG_WARN("rpc proxy truncate table failed", K(ret));
if ((OB_TRY_LOCK_ROW_CONFLICT == ret || OB_TIMEOUT == ret || OB_NOT_MASTER == ret
|| OB_RS_NOT_MASTER == ret || OB_RS_SHUTDOWN == ret || OB_TENANT_NOT_IN_SERVER == ret) && ctx.get_timeout() > 0) {