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

[FEAT MERGE] Implement tenant cloning mechanism

Browse Source 
Co-authored-by: lalalafeier <wufeifei2023@163.com>
Co-authored-by: wanyue-wy <345657357@qq.com>
Co-authored-by: joseph12138 <17862707376@163.com>
This commit is contained in:
obdev
2023-12-14 03:13:29 +00:00
committed by ob-robot
parent b6ee4ddeb7
commit 9187748933
235 changed files with 32253 additions and 1415 deletions
Download Patch File Download Diff File Expand all files Collapse all files

189
src/sql/engine/cmd/ob_clone_executor.cpp Normal file
View File

@ -0,0 +1,189 @@
/**
* Copyright (c) 2021 OceanBase
* OceanBase CE is licensed under Mulan PubL v2.
* You can use this software according to the terms and conditions of the Mulan PubL v2.
* You may obtain a copy of Mulan PubL v2 at:
* http://license.coscl.org.cn/MulanPubL-2.0
* THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND,
* EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT,
* MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE.
* See the Mulan PubL v2 for more details.
*/
#define USING_LOG_PREFIX SQL_ENG
#include "share/ob_common_rpc_proxy.h"
#include "sql/engine/ob_exec_context.h"
#include "sql/engine/cmd/ob_clone_executor.h"
#include "sql/resolver/cmd/ob_tenant_clone_stmt.h"
#include "share/restore/ob_tenant_clone_table_operator.h"
#include "rootserver/restore/ob_tenant_clone_util.h"
namespace oceanbase
{
using namespace common;
using namespace share;
using namespace share::schema;
using namespace rootserver;
namespace sql
{
int ObCloneTenantExecutor::execute(ObExecContext &ctx, ObCloneTenantStmt &stmt)
{
int ret = OB_SUCCESS;
ObTaskExecutorCtx *task_exec_ctx = NULL;
obrpc::ObCommonRpcProxy *common_rpc_proxy = NULL;
const obrpc::ObCloneTenantArg &clone_tenant_arg = stmt.get_clone_tenant_arg();
obrpc::ObCloneTenantRes clone_tenant_res;
const int64_t abs_timeout = ObTimeUtility::current_time() + GCONF._ob_ddl_timeout;
THIS_WORKER.set_timeout_ts(abs_timeout);
if (OB_ISNULL(task_exec_ctx = GET_TASK_EXECUTOR_CTX(ctx))) {
ret = OB_NOT_INIT;
LOG_WARN("get task executor context failed", KR(ret));
} else if (true == stmt.get_if_not_exists()) {
ret = OB_NOT_SUPPORTED;
LOG_WARN("if not exists is true", KR(ret));
LOG_USER_ERROR(OB_NOT_SUPPORTED, "clone tenant with IF NOT EXISTS");
} else if (OB_ISNULL(ctx.get_physical_plan_ctx())) {
ret = OB_ERR_UNEXPECTED;
LOG_WARN("physical plan ctx is null", KR(ret));
} else if (FALSE_IT(ctx.get_physical_plan_ctx()->set_timeout_timestamp(abs_timeout))) {
} else if (OB_ISNULL(common_rpc_proxy = task_exec_ctx->get_common_rpc())) {
ret = OB_NOT_INIT;
LOG_WARN("get common rpc proxy failed", KR(ret));
} else if (!clone_tenant_arg.get_tenant_snapshot_name().empty()) {
// TODO: support tenant snapshot in future
ret = OB_NOT_SUPPORTED;
LOG_WARN("clone tenant with using snapshot clause is not supported", KR(ret), K(clone_tenant_arg));
LOG_USER_ERROR(OB_NOT_SUPPORTED, "clone tenant with using snapshot clause is");
} else if (OB_FAIL(common_rpc_proxy->clone_tenant(clone_tenant_arg, clone_tenant_res))) {
LOG_WARN("rpc proxy clone tenant failed", KR(ret), K(clone_tenant_arg));
} else if (!clone_tenant_res.is_valid()) {
ret = OB_ERR_UNEXPECTED;
LOG_WARN("result is not valid", KR(ret), K(clone_tenant_res));
} else if (OB_FAIL(wait_clone_tenant_finished_(ctx, clone_tenant_res.get_job_id()))) {
LOG_WARN("wait clone tenant finish failed", KR(ret), K(clone_tenant_res));
}
return ret;
}
ERRSIM_POINT_DEF(ERRSIM_WAIT_CLONE_TENANT_FINISHED_ERROR);
int ObCloneTenantExecutor::wait_clone_tenant_finished_(ObExecContext &ctx,
const int64_t job_id)
{
int ret = OB_SUCCESS;
ObCloneJob job;
common::ObMySQLProxy *sql_proxy = nullptr;
ObTimeoutCtx timeout_ctx;
const int64_t trx_timeout = 10 * 60 * 1000 * 1000; // 10min
const int64_t abs_timeout = ObTimeUtility::current_time() + OB_MAX_USER_SPECIFIED_TIMEOUT; // 102 years
THIS_WORKER.set_timeout_ts(abs_timeout);
if (OB_UNLIKELY(ERRSIM_WAIT_CLONE_TENANT_FINISHED_ERROR)) {
ret = ERRSIM_WAIT_CLONE_TENANT_FINISHED_ERROR;
} else if (OB_UNLIKELY(job_id < 0)) {
ret = OB_INVALID_ARGUMENT;
LOG_WARN("invalid argument", KR(ret), K(job_id));
} else if (OB_ISNULL(ctx.get_physical_plan_ctx())) {
ret = OB_ERR_UNEXPECTED;
LOG_WARN("physical plan ctx is null", KR(ret));
} else if (FALSE_IT(ctx.get_physical_plan_ctx()->set_timeout_timestamp(abs_timeout))) {
} else if (OB_FAIL(timeout_ctx.set_trx_timeout_us(trx_timeout))) {
LOG_WARN("failed to set trx timeout us", KR(ret));
} else if (OB_FAIL(timeout_ctx.set_abs_timeout(abs_timeout))) {
LOG_WARN("failed to set abs timeout", KR(ret));
} else if (OB_ISNULL(sql_proxy = ctx.get_sql_proxy())) {
ret = OB_ERR_UNEXPECTED;
LOG_WARN("sql proxy must not be null", KR(ret));
} else {
// if the clone job is successful,
// the according record will be moved to __all_clone_job_history from __all_clone_job;
// if the clone job is failed,
// the according record will be set as failed status in __all_clone_job and
// will be moved to __all_clone_job_history after user executes the "recycle" sql
bool clone_over = false;
while (OB_SUCC(ret) && !clone_over) {
job.reset();
ob_usleep(2 * 1000 * 1000L); // 2s
ObTenantCloneTableOperator table_op;
ObMySQLTransaction trans;
bool exist_in_history = false;
if (THIS_WORKER.is_timeout()) {
ret = OB_TIMEOUT;
LOG_WARN("wait clone tenant timeout", KR(ret), K(job_id));
} else if (OB_FAIL(ctx.check_status())) {
LOG_WARN("check exec ctx failed", KR(ret));
} else if (OB_FAIL(trans.start(sql_proxy, OB_SYS_TENANT_ID))) {
LOG_WARN("fail to start trans", KR(ret));
} else if (OB_FAIL(table_op.init(OB_SYS_TENANT_ID, &trans))) {
LOG_WARN("failed to init table op", KR(ret));
} else if (OB_FAIL(table_op.get_sys_clone_job_history(job_id, job))) {
if (OB_ENTRY_NOT_EXIST == ret) {
ret = OB_SUCCESS;
} else {
LOG_WARN("failed to get clone job history", KR(ret), K(job_id));
}
} else {
exist_in_history = true;
if (job.get_status().is_sys_success_status()) {
clone_over = true;
LOG_INFO("clone tenant successful", K(job));
} else if (job.get_status().is_sys_failed_status()) {
ret = OB_ERR_CLONE_TENANT;
LOG_WARN("clone tenant failed", KR(ret), K(job));
} else {
ret = OB_ERR_UNEXPECTED;
LOG_WARN("unexpected status", KR(ret), K(job));
}
}
if (OB_SUCC(ret) && !exist_in_history) {
int tmp_ret = OB_SUCCESS;
if (OB_FAIL(table_op.get_clone_job_by_job_id(job_id, job))) {
LOG_WARN("failed to get clone job", KR(ret), K(job));
} else if (job.get_status().is_sys_failed_status()) {
ret = OB_ERR_CLONE_TENANT;
LOG_WARN("clone tenant failed", KR(ret), K(job));
} else if (OB_TMP_FAIL(ObTenantCloneUtil::notify_clone_scheduler(OB_SYS_TENANT_ID))) {
// clone job is running
LOG_WARN("notify clone scheduler failed", KR(tmp_ret));
}
}
if (OB_UNLIKELY(OB_TIMEOUT == ret)) {
ret = OB_CLONE_TENANT_TIMEOUT;
LOG_WARN("wait clone tenant timeout", KR(ret), K(job_id));
ObString msg("Please check the details of clone job by DBA_OB_CLONE_PROGRESS or DBA_OB_CLONE_HISTORY");
LOG_USER_ERROR(OB_CLONE_TENANT_TIMEOUT, msg.length(), msg.ptr());
} else if (OB_UNLIKELY(OB_ERR_CLONE_TENANT == ret)) {
int tmp_ret = OB_SUCCESS;
ObArenaAllocator allocator;
ObString err_msg;
if (OB_TMP_FAIL(ObTenantCloneUtil::get_clone_job_failed_message(*sql_proxy,
job_id,
MTL_ID(),
allocator,
err_msg))) {
LOG_WARN("fail to get clone job failed message", KR(ret), K(job_id), K(MTL_ID()));
} else if (!err_msg.empty()) {
LOG_USER_ERROR(OB_ERR_CLONE_TENANT, err_msg.length(), err_msg.ptr());
} else {
ObSqlString format_msg;
ObString failed_status(ObTenantCloneStatus::get_clone_status_str(job.get_status()));
if (OB_TMP_FAIL(format_msg.append_fmt("Tenant clone job failed during the %.*s stage. "
"Please check the details of clone job by "
"DBA_OB_CLONE_PROGRESS or DBA_OB_CLONE_HISTORY",
failed_status.length(), failed_status.ptr()))) {
LOG_WARN("fail to append format", KR(tmp_ret), K(job_id), K(MTL_ID()));
}
LOG_USER_ERROR(OB_ERR_CLONE_TENANT, format_msg.string().length(), format_msg.ptr());
}
}
}
}
return ret;
}
} //end namespace sql
} //end namespace oceanbase