postgresql/lightdb ctas与no space available error

ctas实现逻辑
>    log_newpage    C++ (gdb)
     _bt_blwritepage    C++ (gdb)
     _bt_uppershutdown    C++ (gdb)
     _bt_load    C++ (gdb)
     _bt_leafbuild    C++ (gdb)
     btbuild    C++ (gdb)
     index_build    C++ (gdb)
     index_create    C++ (gdb)
     create_toast_table    C++ (gdb)
     CheckAndCreateToastTable    C++ (gdb)
     NewRelationCreateToastTable    C++ (gdb)
     create_ctas_internal    C++ (gdb)
     intorel_startup    C++ (gdb)
     standard_ExecutorRun    C++ (gdb)
     pgss_ExecutorRun    C++ (gdb)
     ltsa_ExecutorRun    C++ (gdb)
     pgsp_ExecutorRun    C++ (gdb)
     ExecCreateTableAs    C++ (gdb)
     ProcessUtilitySlow    C++ (gdb)
     standard_ProcessUtility    C++ (gdb)
     pgss_ProcessUtility    C++ (gdb)
     ltsa_ProcessUtility    C++ (gdb)
     pgaudit_ProcessUtility_hook    C++ (gdb)
     pg_hint_plan_ProcessUtility    C++ (gdb)
     LtStandbyProcessUtility    C++ (gdb)
     pathman_process_utility_hook    C++ (gdb)
     ProcessUtility    C++ (gdb)
     PortalRunUtility    C++ (gdb)
     PortalRunMulti    C++ (gdb)
     PortalRun    C++ (gdb)
     exec_simple_query    C++ (gdb)
     PostgresMain    C++ (gdb)
     BackendRun    C++ (gdb)
     BackendStartup    C++ (gdb)
     ServerLoop    C++ (gdb)
     PostmasterMain    C++ (gdb)
     main    C++ (gdb)

     libc.so.6!raise    
     libc.so.6!abort    
     errfinish    C++ (gdb)
>    XLogFileInit    C++ (gdb)
     XLogWrite    C++ (gdb)
     XLogFlush    C++ (gdb)
     FlushBuffer    C++ (gdb)
     BufferAlloc    C++ (gdb)
     ReadBuffer_common    C++ (gdb)
     ReadBufferExtended    C++ (gdb)
     ReadBufferBI    C++ (gdb)
     RelationGetBufferForTuple    C++ (gdb)
     heap_insert    C++ (gdb)
     heapam_tuple_insert    C++ (gdb)
     table_tuple_insert    C++ (gdb)
     intorel_receive    C++ (gdb)
     ExecutePlan    C++ (gdb)
     standard_ExecutorRun    C++ (gdb)
     pgss_ExecutorRun    C++ (gdb)
     ltsa_ExecutorRun    C++ (gdb)
     pgsp_ExecutorRun    C++ (gdb)
     ExecCreateTableAs    C++ (gdb)
     ProcessUtilitySlow    C++ (gdb)
     standard_ProcessUtility    C++ (gdb)
     pgss_ProcessUtility    C++ (gdb)
     ltsa_ProcessUtility    C++ (gdb)
     pgaudit_ProcessUtility_hook    C++ (gdb)
     pg_hint_plan_ProcessUtility    C++ (gdb)
     LtStandbyProcessUtility    C++ (gdb)
     pathman_process_utility_hook    C++ (gdb)
     ProcessUtility    C++ (gdb)
     PortalRunUtility    C++ (gdb)
     PortalRunMulti    C++ (gdb)
     PortalRun    C++ (gdb)
     exec_simple_query    C++ (gdb)
     PostgresMain    C++ (gdb)
     BackendRun    C++ (gdb)
     BackendStartup    C++ (gdb)
     ServerLoop    C++ (gdb)
     PostmasterMain    C++ (gdb)
     main    C++ (gdb)

2022-08-06 16:07:51.083299T  @  postmaster  00000[2022-08-06 15:44:05 CST] 0 [165220] LOG:  server process (PID 167657) was terminated by signal 6: Aborted
2022-08-06 16:07:51.083299T  @  postmaster  00000[2022-08-06 15:44:05 CST] 0 [165220] DETAIL:  Failed process was running: create table big_table as select i, 'fewwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww' || i from generate_series(1,10000000) i;
2022-08-06 16:07:51.083339T  @  postmaster  00000[2022-08-06 15:44:05 CST] 0 [165220] LOG:  terminating any other active server processes
2022-08-06 16:07:51.083458T  @  autovacuum launcher  57P02[2022-08-06 15:44:05 CST] 0 [165229] WARNING:  terminating connection because of crash of another server process
2022-08-06 16:07:51.083458T  @  autovacuum launcher  57P02[2022-08-06 15:44:05 CST] 0 [165229] DETAIL:  The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and possibly corrupted shared memory.
2022-08-06 16:07:51.083458T  @  autovacuum launcher  57P02[2022-08-06 15:44:05 CST] 0 [165229] HINT:  In a moment you should be able to reconnect to the database and repeat your command.
2022-08-06 16:07:51.084508T [unknown] zjh@postgres [local] client backend  57P03[2022-08-06 16:07:51 CST] 0 [9389] FATAL:  the database system is in recovery mode
2022-08-06 16:07:51.085146T  @  postmaster  00000[2022-08-06 15:44:05 CST] 0 [165220] LOG:  all server processes terminated; reinitializing
2022-08-06 16:07:51.172925T  @  startup  00000[2022-08-06 16:07:51 CST] 0 [9391] LOG:  database system was interrupted; last known up at 2022-08-06 15:44:05 CST
2022-08-06 16:07:51.184637T  @  startup  00000[2022-08-06 16:07:51 CST] 0 [9391] LOG:  database system was not properly shut down; automatic recovery in progress
2022-08-06 16:07:51.185106T  @  startup  00000[2022-08-06 16:07:51 CST] 0 [9391] LOG:  redo starts at 0/209C9FA0
2022-08-06 16:07:54.077694T  @  startup  00000[2022-08-06 16:07:51 CST] 0 [9391] LOG:  redo done at 0/3FFFFF68
2022-08-06 16:07:54.078338T  @  startup  00000[2022-08-06 16:07:51 CST] 0 [9391] LOG:  checkpoint starting: end-of-recovery immediate
2022-08-06 16:07:55.524716T  @  startup  53100[2022-08-06 16:07:51 CST] 0 [9391] PANIC:  could not write to file "lt_wal/xlogtemp.9391": No space left on device
2022-08-06 16:07:55.526168T  @  postmaster  00000[2022-08-06 15:44:05 CST] 0 [165220] LOG:  startup process (PID 9391) was terminated by signal 6: Aborted
2022-08-06 16:07:55.526177T  @  postmaster  00000[2022-08-06 15:44:05 CST] 0 [165220] LOG:  aborting startup due to startup process failure
2022-08-06 16:07:55.526615T  @  postmaster  00000[2022-08-06 15:44:05 CST] 0 [165220] LOG:  database system is shut down

　　看起来是ctas语句导致了lightdb进程宕机，实际并非如此。业务进程并不会导致lightdb主进程退出，是后台进程写文件失败导致的进程异常退出。lightdb 22.3（已发布）已通过隔离方式实现空间满不会导致主进程宕机。