Docs / Troubleshooting guide

Troubleshooting guide

Errors encountered in real-world operation, and how to handle them.

Each entry shows the sign in the log, the cause and the fix. This list is extended after every incident that is resolved.

Backup tool version does not match the server

Unsupported redo log format / No matching mariabackup version found (Không tìm thấy mariabackup cùng phiên bản)

Cause: mariabackup reads the redo log format of its own version line only. 10.8 and later cannot read a 10.5 redo log; newer versions may also not be backward compatible.

Fix: Install the MariaDB package with the same major.minor as the server on the Agent machine (only the bin folder is needed) and on the Center. Or enter the tool path in step 1 of the Agent / Settings on the Center.

Deadlock while taking the consistency point

Deadlock found when trying to get lock (BACKUP STAGE)

Cause: A known MariaDB 10.5.8 bug when a long transaction is running while the backup changes stage.

Fix: The Agent retries automatically up to 3 times, 30 seconds apart. If it happens often: upgrade MariaDB to the latest 10.5 patch release or to 10.11/11.4. If all tables are InnoDB, no-lock mode is enabled automatically and avoids this error.

Missing backup privileges

Access denied; you need (at least one of) the RELOAD privilege(s)

Cause: The backup account is missing RELOAD, PROCESS, LOCK TABLES, BINLOG MONITOR or SELECT.

Fix: Agent → step 1 → Check connection and privileges (Kiểm tra kết nối và quyền), then run the GRANT statement shown using the root account.

No databases or tables visible

Empty list / catalog 0 tables (Danh sách trống / catalog 0 bảng)

Cause: The user is declared for a different host than the one the Agent connects from (localhost, 127.0.0.1, ::1), or lacks SELECT.

Fix: Check SELECT user, host FROM mysql.user, create an additional user for the correct host, then click Reload list (Đọc lại danh sách) in step 2.

Table in the dictionary but no file

Could not find a valid tablespace file for `db/bang`

Cause: During prepare, mariabackup lists InnoDB tables outside the backup scope (databases not selected). This is completely harmless — the Center collapses them into a single notice.

Fix: No action needed. If the table is within scope and its file really is missing (an orphaned dictionary entry, usually from deleting files by hand): recreate a table with the same name and structure in a temporary database, then DROP TABLE to clean up the dictionary entry, or contact support.

Prepare warns about missing statistics

innodb_table_stats not found / using transient stats

Cause: The backup does not include the mysql database. Version 1.0 always includes mysql; this only happens with backups made from an old configuration.

Fix: Run a new full backup.

Clocks on the two machines differ

clocks differ by N seconds (max 300) (đồng hồ hai máy lệch N giây (tối đa 300))

Cause: Every request is signed with a timestamp; a difference of more than 5 minutes is rejected to prevent replay.

Fix: Enable automatic time synchronization (Windows Time / chrony / systemd-timesyncd) on both machines.

Machine not paired

Request rejected: machine not paired / 401 (Từ chối request: máy chưa ghép cặp / 401)

Cause: The Agent or Center was removed on one side, or a reinstall changed the keys.

Fix: Unpair on the other side, then pair again.

Backup chain broken

Chain broken: #N follows #M but the prepared copy is at #K (Chuỗi gãy: #N nối tiếp #M nhưng bản prepare đang ở #K)

Cause: An incremental in the middle is missing (deleted, failed to send, or the Agent moved to another machine).

Fix: Click Run full now (Chạy full ngay) on the Agent page. Points before the break can still be restored.

Target database already exists

Database `ten` already exists on the Center — choose another name (Database `ten` đã có trên Center — đặt tên khác)

Cause: SaintBackup does not overwrite an existing database.

Fix: Choose another name, or DROP DATABASE it yourself if you are sure it is not needed.

IMPORT TABLESPACE does not work

innodb_file_per_table=0

Cause: The MariaDB on the Center stores all tables in one shared file.

Fix: Set innodb_file_per_table=1 in the Center's my.ini and restart MariaDB.

Disk nearly full

The drive holding the spool on the Agent has … left (Ổ chứa thư mục chờ trên Agent còn …)

Cause: The Center has not received for a long time (turned off, network down), so spooled backups piled up; or the Center repository is full.

Fix: Turn on the Center or fix the connection; reduce the retention days on the Center; move the repository to a larger drive.

MySQL not yet supported

Version 1.0 only supports MariaDB (Bản 1.0 mới hỗ trợ MariaDB)

Cause: Version 1.0 supports MariaDB only.

Fix: Follow the Roadmap.

← PreviousAlerts and logs Next →Security