# params：改过的参数 > 哪些参数不是默认值、在哪设的；改了要重启才生效的报 WARN。 --- LLMS 索引： [llms.txt](/kbdiag-docs/zh/llms.txt) --- 采集于 2026-10-07（UTC+08），本地 KingbaseES V008R006C009B0014 主节点和备节点，数据库 `test`。工具是 Go 版源码提交 `96c874f` 编译的 `~/kbdiag`（linux/amd64 静态二进制）。这是测试环境实测，不代表生产环境验收。WARN 的例子由注入脚本造出来：它用 `ALTER SYSTEM` 改了 `max_files_per_process`（需要重启的参数），reload 了配置，没有重启，所以运行值和文件里的不一样。数值只属于本次采样。 ## 用法 ```text kbdiag params [--json] ``` 没有自己的参数，也没有按名字过滤：看单个参数用 `ksql -c 'show x'`，或者 `--json` 配 jq。连接参数和 [`sessions`](/kbdiag-docs/zh/docs/reference/sessions/) 相同。 ## 主库，正常 ```bash ~/kbdiag params echo EXIT_CODE=$? ``` ```text params OK (KingbaseES V008R006C009B0014, primary, system@local, 2026-10-07T17:20:55+08:00) pending restart: 0 changed: 63 parameters not at their default (sources other than default, override and this connection) this connection sets application_name, default_transaction_read_only, lock_timeout and statement_timeout itself, so their configured values are not shown; per-role and per-database settings show only for this role and database name setting unit source file archive_command export TZ=Asia/Shanghai;/home/kingbase/cluster/install/ki... - configuration file es_rep.conf:10 archive_mode always - configuration file es_rep.conf:9 control_file_copy /home/kingbase/cluster/install/kingbase/copy_file - configuration file es_rep.conf:11 DateStyle ISO, MDY - configuration file kingbase.conf:654 default_text_search_config pg_catalog.english - configuration file kingbase.conf:676 dynamic_shared_memory_type posix - configuration file kingbase.conf:140 fsync on - configuration file es_rep.conf:23 full_page_writes on - configuration file es_rep.conf:3 hot_standby on - configuration file es_rep.conf:13 hot_standby_feedback on - configuration file es_rep.conf:14 lc_messages C - configuration file kingbase.conf:669 lc_monetary en_US.UTF-8 - configuration file kingbase.conf:671 lc_numeric en_US.UTF-8 - configuration file kingbase.conf:672 lc_time en_US.UTF-8 - configuration file kingbase.conf:673 listen_addresses * - configuration file es_rep.conf:1 log_autovacuum_min_duration 0 ms configuration file es_rep.conf:40 log_checkpoints on - configuration file es_rep.conf:38 log_connections on - configuration file es_rep.conf:32 log_destination stderr - configuration file es_rep.conf:44 log_directory sys_log - configuration file kingbase.conf:432 log_disconnections on - configuration file es_rep.conf:33 log_filename kingbase-%d.log - configuration file es_rep.conf:43 logging_collector on - configuration file es_rep.conf:15 log_line_prefix %t [%p]: [%l-1] [%x] user=%u,db=%d,app=%a,client=%h - configuration file es_rep.conf:42 log_lock_waits on - configuration file es_rep.conf:39 log_min_duration_statement 1000 ms configuration file es_rep.conf:47 log_replication_commands on - configuration file es_rep.conf:18 log_rotation_age 1440 min configuration file es_rep.conf:46 log_statement ddl - configuration file es_rep.conf:36 log_temp_files 0 kB configuration file es_rep.conf:41 log_timezone Asia/Shanghai - configuration file kingbase.conf:540 log_truncate_on_rotation on - configuration file es_rep.conf:45 max_connections 100 - configuration file es_rep.conf:7 max_prepared_transactions 100 - configuration file es_rep.conf:21 max_replication_slots 32 - configuration file es_rep.conf:12 max_stack_depth 3072 kB configuration file kingbase.conf:133 max_wal_senders 32 - configuration file es_rep.conf:5 max_wal_size 1024 MB configuration file kingbase.conf:224 min_wal_size 80 MB configuration file kingbase.conf:225 ora_input_emptystr_isnull on - configuration file kingbase.conf:767 ora_integer_div_returnfloat on - configuration file kingbase.conf:768 password_encryption scram-sha-256 - configuration file kingbase.conf:91 port 54321 - configuration file es_rep.conf:2 primary_conninfo user=esrep connect_timeout=10 host=192.168.105.11 port=54... - configuration file kingbase.auto.conf:3 primary_slot_name repmgr_slot_1 - configuration file kingbase.auto.conf:4 shared_buffers 65536 8kB configuration file es_rep.conf:22 shared_preload_libraries repmgr, synonym, plsql, force_view, kdb_flashback,plugin_... - configuration file kingbase.conf:766 synchronous_commit remote_apply - configuration file es_rep.conf:20 synchronous_standby_names ANY 1( node2) - configuration file kingbase.auto.conf:7 sys_stat_statements.track none - configuration file kingbase.auto.conf:5 tcp_keepalives_count 0 - configuration file es_rep.conf:27 tcp_keepalives_idle 0 s configuration file es_rep.conf:25 tcp_keepalives_interval 0 s configuration file es_rep.conf:26 tcp_user_timeout 0 ms configuration file es_rep.conf:28 TimeZone PRC - configuration file es_rep.conf:24 track_real_stats on - configuration file kingbase.auto.conf:6 wal_compression on - configuration file es_rep.conf:19 wal_keep_segments 512 - configuration file es_rep.conf:6 wal_level replica - configuration file es_rep.conf:8 wal_log_hints on - configuration file es_rep.conf:4 wal_receiver_status_interval 2 s configuration file kingbase.conf:323 wal_receiver_timeout 30000 ms configuration file es_rep.conf:30 wal_sender_timeout 30000 ms configuration file es_rep.conf:29 EXIT_CODE=0 ``` - `pending restart` 列出文件里的值和正在运行的值不同的参数（服务器自己的 `pending_restart` 标志）。这里没有。 - `changed` 只列有人设过的参数：来源不是 `default`、`override`（编译或 initdb 定的）以及本连接自己的 `client`/`session`（那是 kbdiag 自己设的，列出来是噪声）。`file` 是设置它的文件和行号。 - `changed` 下面那行说明看不全的两处。kbdiag 连接时自己设了 `application_name`、`default_transaction_read_only`、`lock_timeout`、`statement_timeout`，所以它们的配置值看不到；按角色、按库的设置只看得到当前角色和库的。 - `archive_command` 这类长值会用 `...` 截断；`--json` 里是完整的。 - 备库上的列表类似，`primary_conninfo`、`primary_slot_name`、`synchronous_standby_names` 不同（kes-node2 上实采，同样是 63 个参数）。 ## 没有权限时 `kbdiag_ro`（没有监控角色，走 `127.0.0.1`）： ```bash PGPASSWORD=... ~/kbdiag --host 127.0.0.1 -U kbdiag_ro params echo EXIT_CODE=$? ``` ```text params UNKNOWN (KingbaseES V008R006C009B0014, primary, kbdiag_ro@remote, 2026-10-07T17:21:08+08:00) pending restart: 0 changed: 59 parameters not at their default (sources other than default, override and this connection) this connection sets application_name, default_transaction_read_only, lock_timeout and statement_timeout itself, so their configured values are not shown; per-role and per-database settings show only for this role and database name setting unit source file archive_command export TZ=Asia/Shanghai;/home/kingbase/cluster/install/ki... - configuration file ? archive_mode always - configuration file ? control_file_copy /home/kingbase/cluster/install/kingbase/copy_file - configuration file ? DateStyle ISO, MDY - configuration file ? default_text_search_config pg_catalog.english - configuration file ? dynamic_shared_memory_type posix - configuration file ? fsync on - configuration file ? full_page_writes on - configuration file ? hot_standby on - configuration file ? hot_standby_feedback on - configuration file ? lc_messages C - configuration file ? lc_monetary en_US.UTF-8 - configuration file ? lc_numeric en_US.UTF-8 - configuration file ? lc_time en_US.UTF-8 - configuration file ? listen_addresses * - configuration file ? log_autovacuum_min_duration 0 ms configuration file ? log_checkpoints on - configuration file ? log_connections on - configuration file ? log_destination stderr - configuration file ? log_disconnections on - configuration file ? logging_collector on - configuration file ? log_line_prefix %t [%p]: [%l-1] [%x] user=%u,db=%d,app=%a,client=%h - configuration file ? log_lock_waits on - configuration file ? log_min_duration_statement 1000 ms configuration file ? log_replication_commands on - configuration file ? log_rotation_age 1440 min configuration file ? log_statement ddl - configuration file ? log_temp_files 0 kB configuration file ? log_timezone Asia/Shanghai - configuration file ? log_truncate_on_rotation on - configuration file ? max_connections 100 - configuration file ? max_prepared_transactions 100 - configuration file ? max_replication_slots 32 - configuration file ? max_stack_depth 3072 kB configuration file ? max_wal_senders 32 - configuration file ? max_wal_size 1024 MB configuration file ? min_wal_size 80 MB configuration file ? ora_input_emptystr_isnull on - configuration file ? ora_integer_div_returnfloat on - configuration file ? password_encryption scram-sha-256 - configuration file ? port 54321 - configuration file ? primary_slot_name repmgr_slot_1 - configuration file ? shared_buffers 65536 8kB configuration file ? synchronous_commit remote_apply - configuration file ? synchronous_standby_names ANY 1( node2) - configuration file ? sys_stat_statements.track none - configuration file ? tcp_keepalives_count 3 - configuration file ? tcp_keepalives_idle 2 s configuration file ? tcp_keepalives_interval 2 s configuration file ? tcp_user_timeout 9000 ms configuration file ? TimeZone PRC - configuration file ? track_real_stats on - configuration file ? wal_compression on - configuration file ? wal_keep_segments 512 - configuration file ? wal_level replica - configuration file ? wal_log_hints on - configuration file ? wal_receiver_status_interval 2 s configuration file ? wal_receiver_timeout 30000 ms configuration file ? wal_sender_timeout 30000 ms configuration file ? only the parameters this account may read: superuser-only ones are not listed redacted: 59 rows of params.changed hide sourcefile, sourceline (insufficient_privilege; grant sys_monitor) EXIT_CODE=3 ``` - `file` 是 `?`：这个账号看不到 `sourcefile` 和 `sourceline`，记在 `redacted` 里。 - 列表更短（59 个，不是 63 个），因为超级用户专属的参数这个账号根本看不到，最后一行说明了这点。这些参数的 `pending_restart` 也就看不到，所以没有 finding 时结论是 UNKNOWN：看不到的东西，kbdiag 不说 OK。 ## 改了还没生效 ```bash ~/kbdiag params echo EXIT_CODE=$? ``` ```text params WARN (KingbaseES V008R006C009B0014, primary, system@local, 2026-10-07T17:21:15+08:00) [WARN] params.pending_restart parameter max_files_per_process was changed but the instance still runs with 1000: the new value takes effect only after a restart, the next one included verify: select sourcefile, sourceline, setting, applied, error from sys_file_settings where name = 'max_files_per_process' # the value waiting in the file (needs a superuser) fix: ALTER SYSTEM SET max_files_per_process = '1000' # if the change was not meant: writes the running value back; after SELECT sys_reload_conf() the flag clears (a reload clears it only when the file's value equals the running one, so ALTER SYSTEM RESET alone leaves it until a restart). Otherwise plan the restart pending restart: 1 name running file max_files_per_process 1000 - changed: 64 parameters not at their default (sources other than default, override and this connection) this connection sets application_name, default_transaction_read_only, lock_timeout and statement_timeout itself, so their configured values are not shown; per-role and per-database settings show only for this role and database name setting unit source file archive_command export TZ=Asia/Shanghai;/home/kingbase/cluster/install/ki... - configuration file es_rep.conf:10 archive_mode always - configuration file es_rep.conf:9 control_file_copy /home/kingbase/cluster/install/kingbase/copy_file - configuration file es_rep.conf:11 DateStyle ISO, MDY - configuration file kingbase.conf:654 default_text_search_config pg_catalog.english - configuration file kingbase.conf:676 dynamic_shared_memory_type posix - configuration file kingbase.conf:140 fsync on - configuration file es_rep.conf:23 full_page_writes on - configuration file es_rep.conf:3 hot_standby on - configuration file es_rep.conf:13 hot_standby_feedback on - configuration file es_rep.conf:14 lc_messages C - configuration file kingbase.conf:669 lc_monetary en_US.UTF-8 - configuration file kingbase.conf:671 lc_numeric en_US.UTF-8 - configuration file kingbase.conf:672 lc_time en_US.UTF-8 - configuration file kingbase.conf:673 listen_addresses * - configuration file es_rep.conf:1 log_autovacuum_min_duration 0 ms configuration file es_rep.conf:40 log_checkpoints on - configuration file es_rep.conf:38 log_connections on - configuration file es_rep.conf:32 log_destination stderr - configuration file es_rep.conf:44 log_directory sys_log - configuration file kingbase.conf:432 log_disconnections on - configuration file es_rep.conf:33 log_filename kingbase-%d.log - configuration file es_rep.conf:43 logging_collector on - configuration file es_rep.conf:15 log_line_prefix %t [%p]: [%l-1] [%x] user=%u,db=%d,app=%a,client=%h - configuration file es_rep.conf:42 log_lock_waits on - configuration file es_rep.conf:39 log_min_duration_statement 1000 ms configuration file es_rep.conf:47 log_replication_commands on - configuration file es_rep.conf:18 log_rotation_age 1440 min configuration file es_rep.conf:46 log_statement ddl - configuration file es_rep.conf:36 log_temp_files 0 kB configuration file es_rep.conf:41 log_timezone Asia/Shanghai - configuration file kingbase.conf:540 log_truncate_on_rotation on - configuration file es_rep.conf:45 max_connections 100 - configuration file es_rep.conf:7 max_files_per_process 1000 - default - max_prepared_transactions 100 - configuration file es_rep.conf:21 max_replication_slots 32 - configuration file es_rep.conf:12 max_stack_depth 3072 kB configuration file kingbase.conf:133 max_wal_senders 32 - configuration file es_rep.conf:5 max_wal_size 1024 MB configuration file kingbase.conf:224 min_wal_size 80 MB configuration file kingbase.conf:225 ora_input_emptystr_isnull on - configuration file kingbase.conf:767 ora_integer_div_returnfloat on - configuration file kingbase.conf:768 password_encryption scram-sha-256 - configuration file kingbase.conf:91 port 54321 - configuration file es_rep.conf:2 primary_conninfo user=esrep connect_timeout=10 host=192.168.105.11 port=54... - configuration file kingbase.auto.conf:3 primary_slot_name repmgr_slot_1 - configuration file kingbase.auto.conf:4 shared_buffers 65536 8kB configuration file es_rep.conf:22 shared_preload_libraries repmgr, synonym, plsql, force_view, kdb_flashback,plugin_... - configuration file kingbase.conf:766 synchronous_commit remote_apply - configuration file es_rep.conf:20 synchronous_standby_names ANY 1( node2) - configuration file kingbase.auto.conf:7 sys_stat_statements.track none - configuration file kingbase.auto.conf:5 tcp_keepalives_count 0 - configuration file es_rep.conf:27 tcp_keepalives_idle 0 s configuration file es_rep.conf:25 tcp_keepalives_interval 0 s configuration file es_rep.conf:26 tcp_user_timeout 0 ms configuration file es_rep.conf:28 TimeZone PRC - configuration file es_rep.conf:24 track_real_stats on - configuration file kingbase.auto.conf:6 wal_compression on - configuration file es_rep.conf:19 wal_keep_segments 512 - configuration file es_rep.conf:6 wal_level replica - configuration file es_rep.conf:8 wal_log_hints on - configuration file es_rep.conf:4 wal_receiver_status_interval 2 s configuration file kingbase.conf:323 wal_receiver_timeout 30000 ms configuration file es_rep.conf:30 wal_sender_timeout 30000 ms configuration file es_rep.conf:29 EXIT_CODE=1 ``` - `params.pending_restart` 是 WARN，每个参数一条，线就是服务器自己的 `pending_restart` 标志。现在业务没受影响：实例还在跑旧值。但下一次重启（包括故障切换后的重启）会突然换成新值。 - `pending restart` 里显示的是运行值；文件里等着的值这里显示 `-`。`verify` 给了一条查询（`sys_file_settings`，需要超级用户）。 - `fix` 是把运行值写回去：`ALTER SYSTEM SET x = '<运行值>'`，再 `SELECT sys_reload_conf()`。PostgreSQL 12 的 reload 只在文件里的值等于运行值时才清掉 pending 标志，只做 `ALTER SYSTEM RESET` 会一直挂到下次重启。如果这个改动是有意的，就该去安排重启。 - `max_files_per_process` 在 `changed` 里来源是 `default`：等着重启的参数不管来源都会列出，因为改了之后它的来源仍然读成 `default`。 - 配置文件有错（`sys_file_settings.error`）这一项没做：`kbdiag_ro` 读不了那个视图。 ## 退出码 | 退出码 | 含义 | |---|---| | 0 | OK | | 1 | WARN：有参数改了要重启才生效 | | 3 | UNKNOWN：有数据没采到或看不到 | | 64 | 参数错误 | | 69 | 连不上数据库 | [返回使用手册](/kbdiag-docs/zh/docs/reference/)