add fe meta editlog skip config

docs/admin-manual/config/fe-config.md

@@ -2738,6 +2738,22 @@ Default value: 1
 
 Used to control the interval at which ProfileManager performs profile garbage collection. During garbage collection, ProfileManager purges excess and expired profiles from memory and disk to save memory.
 
+#### `force_skip_journal_ids`
+
+Default value: {}
+
+When FE restarts, if the replay process encounters corrupted journal entries during metadata image replay that cause the replay thread to crash, preventing FE from starting normally, you can configure the problematic journal IDs to temporarily skip these corrupted entries, allowing FE to start successfully. Default value is an empty set {}.
+
+**Note:** This configuration is intended for emergency recovery scenarios only. Skipping corrupted journal entries may lead to metadata inconsistency. It is recommended to clear this configuration after the underlying issue is resolved.
+
+#### `skip_operation_types_on_replay_exception`
+
+Default value: {-1, -1}
+
+When FE replays edit logs (editlog), if specific operation types encounter exceptions that prevent FE from starting, you can configure the edit log operation type enumeration values to be ignored, allowing the replay thread to skip these problematic operations and continue processing other log entries. Default value is {-1, -1}.
+
+**Note:** This configuration is also intended for emergency recovery scenarios only. Skipping specific operation types may affect metadata integrity. It is recommended to clear this configuration after the underlying issue is resolved.
+
 ### Compute and Storage Disaggregated Mode
 
 #### `cluster_id`

i18n/zh-CN/docusaurus-plugin-content-docs/current/admin-manual/config/fe-config.md

@@ -2727,6 +2727,22 @@ Doris 为了兼用 mysql 周边工具生态，会内置一个名为 mysql 的数
 
 对于自动分区表，防止用户意外创建大量分区，每个 OLAP 表允许的分区数量为`max_auto_partition_num`。默认 2000。
 
+#### `force_skip_journal_ids`
+
+默认值：{}
+
+当 FE 重启时，如果回放元数据镜像过程中遇到异常的日志条目导致回放线程异常退出，从而无法正常启动 FE 时，可以通过配置异常的日志 ID 来临时跳过这些有问题的日志条目，使 FE 能够正常启动。默认值为空集合 {}。
+
+**注意：** 此配置仅用于紧急情况下的临时恢复，跳过异常日志可能会导致元数据不一致，建议在问题解决后及时清理此配置。
+
+#### `skip_operation_types_on_replay_exception`
+
+默认值：{-1, -1}
+
+当 FE 回放日志（editlog）时，如果遇到特定操作类型的异常导致 FE 无法启动，可以通过配置需要忽略的日志操作类型枚举值来跳过这些异常操作，使回放线程能够继续处理其他日志条目。默认值为 {-1, -1}。
+
+**注意：** 此配置同样仅用于紧急恢复场景，跳过特定操作类型可能会影响元数据的完整性，建议在问题解决后及时清理此配置。
+
 ### 计算与存储分离模式
 
 #### `cluster_id`

i18n/zh-CN/docusaurus-plugin-content-docs/version-3.0/admin-manual/config/fe-config.md

@@ -2729,6 +2729,22 @@ Doris 为了兼用 mysql 周边工具生态，会内置一个名为 mysql 的数
 
 用于控制 ProfileManager 进行 Profile 垃圾回收的间隔时间，垃圾回收期间 ProfileManager 会把多余的以及过期的 profile 从内存和磁盘中清理掉，节省内存。
 
+#### `force_skip_journal_ids`
+
+默认值：{}
+
+当 FE 重启时，如果回放元数据镜像过程中遇到异常的日志条目导致回放线程异常退出，从而无法正常启动 FE 时，可以通过配置异常的日志 ID 来临时跳过这些有问题的日志条目，使 FE 能够正常启动。默认值为空集合 {}。
+
+**注意：** 此配置仅用于紧急情况下的临时恢复，跳过异常日志可能会导致元数据不一致，建议在问题解决后及时清理此配置。
+
+#### `skip_operation_types_on_replay_exception`
+
+默认值：{-1, -1}
+
+当 FE 回放日志（editlog）时，如果遇到特定操作类型的异常导致 FE 无法启动，可以通过配置需要忽略的日志操作类型枚举值来跳过这些异常操作，使回放线程能够继续处理其他日志条目。默认值为 {-1, -1}。
+
+**注意：** 此配置同样仅用于紧急恢复场景，跳过特定操作类型可能会影响元数据的完整性，建议在问题解决后及时清理此配置。
+
 ### 存算分离模式
 
 #### `cluster_id`

versioned_docs/version-2.1/admin-manual/config/fe-config.md

@@ -2753,3 +2753,19 @@ To ensure compatibility with the MySQL ecosystem, Doris includes a built-in data
 Default value: 2000
 
 For auto-partitioned tables to prevent users from accidentally creating a large number of partitions, the number of partitions allowed per OLAP table is `max_auto_partition_num`. Default 2000.
+
+#### `force_skip_journal_ids`
+
+Default value: {}
+
+When FE restarts, if the replay process encounters corrupted journal entries during metadata image replay that cause the replay thread to crash, preventing FE from starting normally, you can configure the problematic journal IDs to temporarily skip these corrupted entries, allowing FE to start successfully. Default value is an empty set {}.
+
+**Note:** This configuration is intended for emergency recovery scenarios only. Skipping corrupted journal entries may lead to metadata inconsistency. It is recommended to clear this configuration after the underlying issue is resolved.
+
+#### `skip_operation_types_on_replay_exception`
+
+Default value: {-1, -1}
+
+When FE replays edit logs (editlog), if specific operation types encounter exceptions that prevent FE from starting, you can configure the edit log operation type enumeration values to be ignored, allowing the replay thread to skip these problematic operations and continue processing other log entries. Default value is {-1, -1}.
+
+**Note:** This configuration is also intended for emergency recovery scenarios only. Skipping specific operation types may affect metadata integrity. It is recommended to clear this configuration after the underlying issue is resolved.

versioned_docs/version-3.0/admin-manual/config/fe-config.md

@@ -2730,6 +2730,22 @@ Default value: 1
 
 Used to control the interval at which ProfileManager performs profile garbage collection. During garbage collection, ProfileManager purges excess and expired profiles from memory and disk to save memory.
 
+#### `force_skip_journal_ids`
+
+Default value: {}
+
+When FE restarts, if the replay process encounters corrupted journal entries during metadata image replay that cause the replay thread to crash, preventing FE from starting normally, you can configure the problematic journal IDs to temporarily skip these corrupted entries, allowing FE to start successfully. Default value is an empty set {}.
+
+**Note:** This configuration is intended for emergency recovery scenarios only. Skipping corrupted journal entries may lead to metadata inconsistency. It is recommended to clear this configuration after the underlying issue is resolved.
+
+#### `skip_operation_types_on_replay_exception`
+
+Default value: {-1, -1}
+
+When FE replays edit logs (editlog), if specific operation types encounter exceptions that prevent FE from starting, you can configure the edit log operation type enumeration values to be ignored, allowing the replay thread to skip these problematic operations and continue processing other log entries. Default value is {-1, -1}.
+
+**Note:** This configuration is also intended for emergency recovery scenarios only. Skipping specific operation types may affect metadata integrity. It is recommended to clear this configuration after the underlying issue is resolved.
+
 ### Compute and Storage Disaggregated Mode
 
 #### `cluster_id`