add index task state show cancelled but actually succeed when inject ddl owner network partition with global sort

## Bug Report

Please answer these questions before submitting your issue. Thanks!

### 1. Minimal reproduce step (Required)
tidb_enable_dist_task='on' and enable global sort
1、run workload
2、add index for one table
3、 inject ddl owner network partition

### 2. What did you expect to see? (Required)
add index task success

### 3. What did you see instead (Required)
add index task state show cancelled but actually succeed when inject ddl owner network partition

the status of ddl job is not synced or running (now: 2023-11-16 06:18:39, jobId: 600, job type: add index /* ingest cloud */, state: cancelled)
operatorLogs:
[2023-11-16 06:13:37] ###### start adding index
alter table sbtest1 add index index_test_1700086417962 (c)
[2023-11-16 06:13:37] ###### wait for ddl job finish
[2023-11-16 06:18:39] ###### wait for import job to finish failed
select job_id, job_type, state from information_schema.ddl_jobs where query = 'alter table sbtest1 add index index_test_1700086417962 (c)'
jobId: 600, job type: add index /* ingest cloud */, state: cancelled

### 4. What is your TiDB version? (Required)
git hash：151b5b1df9d595bf0b568ff82f2ea2e13ae3eddb