[BUG] WARNING It was not possible to process file: because real_db_name isn't found.

[bmesich]

Describe the bug
When attempting to restore a database, I am getting the following warning:

** (myloader:124446): WARNING **: 11:38:56.769: It was not possible to process file: sitemanager.tutorials_tbl-metadata (2) because real_db_name isn't found. We might renqueue it, take into account that restores without schema-create files are not supported

This is very similar to the issue 683 and appears to have been fixed in 708. In this case, a WARNING is being generated rather than a CRITICAL log severity.

To Reproduce
I used more-or-less the same steps to reproduce the warning as was shown in issue 683

MySQL Console

create database sitemanager;
use sitemanager;
create table tutorials_tbl(
tutorial_id INT NOT NULL AUTO_INCREMENT,
tutorial_title VARCHAR(100) NOT NULL,
tutorial_author VARCHAR(40) NOT NULL,
submission_date DATE,
PRIMARY KEY ( tutorial_id )
);

BASH Shell

mydumper --routines --triggers -S /var/lib/mysql/mysql.sock -o /backup-prep/sitemanager -B sitemanager
myloader -S /var/lib/mysql/mysql.sock -d /backup-prep/sitemanager

myloader generates the following WARNING:

** (myloader:124446): WARNING **: 11:38:56.769: It was not possible to process file: sitemanager.tutorials_tbl-metadata (2) because real_db_name isn't found. We might renqueue it, take into account that restores without schema-create files are not supported

What mydumper and myloader version has been used?
myloader 0.12.7-3, built against MySQL 10.6.9
mydumper 0.12.7-3, built against MySQL 10.6.9-MariaDB

Expected behavior
It would appear that the database is restored successfully if the WARNING log messages are ignored. Perhaps this WARNING should be suppressed or prevented from appearing.

Log

** Message: 12:02:07.520: Server version reported as: 10.6.9-MariaDB-log
** Message: 12:02:07.523: Intermediate queue ending

** (myloader:125010): WARNING **: 12:02:07.524: It was not possible to process file: sitemanager.tutorials_tbl-metadata (2) because real_db_name isn't found. We might renqueue it, take into account that restores without schema-create files are not supported
** Message: 12:02:07.524: Thread 1 restoring create database on `sitemanager` from sitemanager-schema-create.sql
** Message: 12:02:07.524: Intermediate thread ended
** Message: 12:02:07.524: Intermediate thread ended
** Message: 12:02:07.524: Thread 1: Data import ended
** Message: 12:02:07.524: Thread 3: Data import ended
** Message: 12:02:07.524: Thread 3: Starting post import task over table
** Message: 12:02:07.524: Thread 2 restoring table `sitemanager`.`tutorials_tbl` from //backup-prep/sitemanager/sitemanager.tutorials_tbl-schema.sql
** Message: 12:02:07.524: Thread 2: Creating table `sitemanager`.`tutorials_tbl` from content in //backup-prep/sitemanager/sitemanager.tutorials_tbl-schema.sql
** Message: 12:02:07.524: Thread 1: Starting post import task over table
** Message: 12:02:07.524: Thread 4: Data import ended
** Message: 12:02:07.524: Thread 4: Starting post import task over table
** Message: 12:02:07.530: Thread 2: Data import ended
** Message: 12:02:07.530: Thread 2: Starting post import task over table
** Message: 12:02:07.531: Starting table checksum verification

Backup

-rw-r--r-- 1 root root  75 Sep 23 11:33 metadata
-rw-r--r-- 1 root root  98 Sep 23 11:33 sitemanager-schema-create.sql
-rw-r--r-- 1 root root   1 Sep 23 11:33 sitemanager.tutorials_tbl-metadata
-rw-r--r-- 1 root root 377 Sep 23 11:33 sitemanager.tutorials_tbl-schema.sql

How to repeat
See steps above to repeat. Note that using "-B ${DB_NAME}" argument to myloader will suppress the log WARNING.

Environment (please complete the following information):

• OS version: RHEL9.0
• MyDumper version: mydumper 0.12.7-3, built against MySQL 10.6.9-MariaDB

[davidducos]

I think that we don't print any message if a metadata file is processed, so we don't know if the file was processed but I think that we are successfully processing those files.
I'm agree that we can remove the Warning, we can move to debug and just print a warning or critical if a file was not processed because it reaches to the maximal amount of attempts which is 5.

[fenidik]

Have same issue. You should test with data. Because when I'm trying to restore it creates schema but not uploading the data.

mydumper --compress --defaults-file .my.cnf -t 10 # defaults file only contains db credentials
myloader --defaults-file ../.my.cnf -t 10 -d .

[davidducos]

@fenidik, we are testing with data. Can you give more details about your issue? MyDumper version used? Database version? a test case with CREATE TABLE and INSERT statements?

[bigkahuna1986]

Also have the same issue as @fenidik. myloader is producing a huge amount of warnings but not restoring any data.

$ myloader --version
myloader 0.12.7-3, built against MySQL 8.0.29-21

Server version: 10.6.10-MariaDB-1:10.6.10+maria~ubu2004-log mariadb.org binary distribution

My workflow is I restore our production database to my workstation from a mysqldump. We have about 100 databases most of which are less than 100mb (total size is around 15gb, biggest DB is just over 1gb). At this point I launch our application and run tests with no problems so I know the restore worked. Then I create a dump file using the following command:
$ mydumper --defaults-file mydumper.conf --regex 'my_dbs_prefix*' --outputdir /backups --compress

It's worth noting Maria is in a docker container.

The dump seems to work well, with the /backups directory containing ~625mb of gzipped data.

mydumper.conf contains:
[mydumper]
user = {redacted}
password = {redacted}
socket = /opt/mariadb/var/mysql.sock

[myloader]
user = {redacted}
password = {redacted}
socket = /opt/mariadb/var/mysql.sock
innodb-optimize-keys = AFTER_IMPORT_PER_TABLE

Since this is production data I can't upload that as a test case. I'll try to find something smaller

Load command:
myloader --directory /backups --defaults-file mydumper.conf --queries-per-transaction 5000 --threads 8 --overwrite-tables

At this point, myloader seems to create begin overwriting the first table but never does the work. For about 90 seconds Maria has moderate CPU usage (~30-50%), then goes completely idle. myloader is using 100% CPU, Maria is idle and can be restarted. I have to kill -9 the process.

EDIT
I removed the innodb-optimize-keys = AFTER_IMPORT_PER_TABLE option from the conf, and it now finishes in 8 minutes. Still has a ton of warning messages as per the OP.

[davidducos]

Hi @bigkahuna1986,
Can you test latest prerelease (v0.13.0-6) and let me know if you see your issue with it?

[lizhongxuan]

I get the same error. When backing up a database, restore cannot use its original name. Otherwise, an error will be reported.

backup: mydumper -h 127.0.0.1 -u root -p abc123 -P 3306 -B testdb -r 1000000 -c -o ./

error recovery: myloader -h 127.0.0.1 -u root -p abc123 -P 3306 -B testdb -o -d ./

true recovery : myloader -h 127.0.0.1 -u root -p abc123 -P 3306 -o -d ./

[davidducos]

@Vega-Punk what version are you using?

[lizhongxuan]

@Vega-Punk你用的是什么版本？

mydumper 0.12.3-3, built against MySQL 5.7.37-40
myloader 0.12.3-3, built against MySQL 5.7.37-40

centos7.6 3.10.0-693.el7.x86_64

[davidducos]

Please, test with the latest prerelease.