Merge pull request #7 from woblerr/history-clean

Add history-clean command.
woblerr · Feb 8, 2024 · d4e1ea2 · d4e1ea2
2 parents b5ca7db + 448519e
commit d4e1ea2
Show file tree

Hide file tree

Showing 23 changed files with 917 additions and 100 deletions.
diff --git a/COMMANDS.md b/COMMANDS.md
@@ -12,14 +12,19 @@
 - [Display information about backups (`backup-info`)](#display-information-about-backups-backup-info)
   - [Examples](#examples-2)
   - [Using container](#using-container-2)
-- [Migrate history database (`history-migrate`)](#migrate-history-database-history-migrate)
+- [Clean failed and deleted backups from the history database (`history-clean`)](#clean-failed-and-deleted-backups-from-the-history-database-history-clean)
   - [Examples](#examples-3)
+    - [Delete information about failed and deleted backups from history database older than n days](#delete-information-about-failed-and-deleted-backups-from-history-database-older-than-n-days)
+    - [Delete all backups using storage plugin older than timestamp](#delete-all-backups-using-storage-plugin-older-than-timestamp-1)
   - [Using container](#using-container-3)
-- [Display the report for a specific backup (`report-info`)](#display-the-report-for-a-specific-backup-report-info)
+- [Migrate history database (`history-migrate`)](#migrate-history-database-history-migrate)
   - [Examples](#examples-4)
+  - [Using container](#using-container-4)
+- [Display the report for a specific backup (`report-info`)](#display-the-report-for-a-specific-backup-report-info)
+  - [Examples](#examples-5)
     - [Display the backup report from local storage](#display-the-backup-report-from-local-storage)
     - [Display the backup report using storage plugin](#display-the-backup-report-using-storage-plugin)
-  - [Using container](#using-container-4)
+  - [Using container](#using-container-5)
 
 # Delete all existing backups older than the specified time condition (`backup-clean`)
 
@@ -82,7 +87,7 @@ gpBackMan returns a message:
 Delete all backups older than 7 days and all dependent backups:
 ```bash
 ./gpbackman backup-clean \
-  --older-than-day 7 \
+  --older-than-days 7 \
   --plugin-config /tmp/gpbackup_plugin_config.yaml \
   --cascade
 ```
@@ -98,7 +103,7 @@ Delete all backups older than timestamp `20240101100000` and all dependent backu
 
 ## Using container
 
-Delete all backups using `gpbackup_s3_plugin` storage plugin: older than 7 days:
+Delete all backups using `gpbackup_s3_plugin` storage plugin older than 7 days:
 ```bash
 docker run \
   --name gpbackman \
@@ -109,7 +114,7 @@ docker run \
   -v /path/to/gpbackup_plugin_config.yaml:/tmp/gpbackup_plugin_config.yaml \
   gpbackman \
   gpbackman backup-clean \
-  --older-than-day 7 \
+  --older-than-days 7 \
   --history-db /data/master/gpseg-1/gpbackup_history.db \
   --plugin-config /tmp/gpbackup_plugin_config.yaml
 ```
@@ -215,9 +220,9 @@ Display information about backups.
 
 By default, only active backups or backups with deletion status "In progress" from gpbackup_history.db are displayed.
 
-To additional display deleted backups, use the --show-deleted option.
-To additional display failed backups, use the --show-failed option.
-To display all backups, use --show-deleted  and --show-failed options together.
+To additional display deleted backups, use the --deleted option.
+To additional display failed backups, use the --failed option.
+To display all backups, use --deleted and --failed options together.
 
 To display backups of a specific type, use the --type option.
 
@@ -235,10 +240,10 @@ Usage:
   gpbackman backup-info [flags]
 
 Flags:
-  -h, --help           help for backup-info
-      --show-deleted   show deleted backups
-      --show-failed    show failed backups
-      --type string    backup type filter (full, incremental, data-only, metadata-only)
+      --deleted       show deleted backups
+      --failed        show failed backups
+  -h, --help          help for backup-info
+      --type string   backup type filter (full, incremental, data-only, metadata-only)
 
 Global Flags:
       --history-db string          full path to the gpbackup_history.db file
@@ -295,8 +300,8 @@ Display info for active backups from `gpbackup_history.db`:
 Display info for all backups from `gpbackup_history.yaml`:
 ```bash
 ./gpbackman backup-info \
-  --show-deleted \
-  --show-failed \
+  --deleted \
+  --failed \
   --history-file /data/master/gpseg-1/gpbackup_history.yaml
  TIMESTAMP      | DATE                     | STATUS  | DATABASE | TYPE          | OBJECT FILTERING | PLUGIN             | DURATION | DATE DELETED 
 ----------------+--------------------------+---------+----------+---------------+------------------+--------------------+----------+--------------
@@ -321,6 +326,84 @@ docker run \
   --history-db /data/master/gpseg-1/gpbackup_history.db
 ```
 
+# Clean failed and deleted backups from the history database (`history-clean`)
+
+Available options for `history-clean` command and their description:
+
+```bash
+./gpbackman history-clean -h
+
+Clean failed and deleted backups from the history database.
+Only the database is being cleaned up.
+
+By default, information is deleted only about failed backups from gpbackup_history.db.
+
+To delete information about deleted backups, use the --deleted option.
+
+To delete information about backups older than the given timestamp, use the --before-timestamp option. 
+To delete information about backups older than the given number of days, use the --older-than-day option. 
+Only --older-than-days or --before-timestamp option must be specified, not both.
+
+The gpbackup_history.db file location can be set using the --history-db option.
+Can be specified only once. The full path to the file is required.
+
+The gpbackup_history.yaml file location can be set using the --history-file option.
+Can be specified multiple times. The full path to the file is required.
+
+If no --history-file or --history-db options are specified, the history database will be searched in the current directory.
+
+Only --history-file or --history-db option can be specified, not both.
+
+Usage:
+  gpbackman history-clean [flags]
+
+Flags:
+      --before-timestamp string   delete information about backups older than the given timestamp
+      --deleted                   delete information about deleted backups
+  -h, --help                      help for history-clean
+      --older-than-days uint      delete information about backups older than the given number of days
+
+Global Flags:
+      --history-db string          full path to the gpbackup_history.db file
+      --history-file stringArray   full path to the gpbackup_history.yaml file, could be specified multiple times
+      --log-file string            full path to log file directory, if not specified, the log file will be created in the $HOME/gpAdminLogs directory
+      --log-level-console string   level for console logging (error, info, debug, verbose) (default "info")
+      --log-level-file string      level for file logging (error, info, debug, verbose) (default "info")
+```
+
+## Examples
+### Delete information about failed and deleted backups from history database older than n days
+Delete information about failed and deleted backups from history database older than 7 days:
+```bash
+./gpbackman history-clean \
+  --older-than-days 7 \
+  --deleted
+```
+
+### Delete all backups using storage plugin older than timestamp
+Delete information about failed and deleted backups from history database older than timestamp `20240101100000`:
+```bash
+./gpbackman history-clean \
+  --before-timestamp 20240101100000 \
+  --deleted
+```
+
+## Using container
+
+Delete information about failed and deleted backups from history database older than 7 days:
+```bash
+docker run \
+  --name gpbackman \
+  -e GPBACKMAN_UID=$(id -u) \
+  -e GPBACKMAN_GID=$(id -g) \
+  -v /data/master/gpseg-1/gpbackup_history.db:/data/master/gpseg-1/gpbackup_history.db \
+  gpbackman \
+  gpbackman history-clean \
+  --older-than-days 7 \
+  --history-db /data/master/gpseg-1/gpbackup_history.db \
+  --deleted
+```
+
 # Migrate history database (`history-migrate`)
 
 Available options for `history-migrate` command and their description:

diff --git a/Makefile b/Makefile
@@ -21,6 +21,7 @@ test-e2e:
 	@make test-e2e_report-info
 	@make test-e2e_backup-delete
 	@make test-e2e_history-migrate
+	@make test-e2e_history-clean
 
 .PHONY: test-e2e_backup-info
 test-e2e_backup-info:
@@ -43,6 +44,13 @@ test-e2e_backup-clean:
 	$(call run_docker_compose,backup-clean)
 	$(call down_docker_compose)
 
+.PHONY: test-e2e_history-clean
+test-e2e_history-clean:
+	@echo "Run end-to-end tests for $(APP_NAME) for history-clean command"
+	$(call down_docker_compose)
+	$(call run_docker_compose,history-clean)
+	$(call down_docker_compose)
+
 .PHONY: test-e2e_history-migrate
 test-e2e_history-migrate:
 	@echo "Run end-to-end tests for $(APP_NAME) for history-migrate command"

diff --git a/README.md b/README.md
@@ -14,6 +14,7 @@ The utility works with both history database formats: `gpbackup_history.yaml` fi
 * display the backup report for existing backups;
 * delete existing backups from local storage or using storage plugins (for example, [S3 Storage Plugin](https://github.com/greenplum-db/gpbackup-s3-plugin));
 * delete all existing backups from local storage or using storage plugins older than the specified time condition;
+* clean failed and deleted backups from the history database;
 * migrate history database from `gpbackup_history.yaml` format to `gpbackup_history.db` SQLite format.
 
 ## Commands
@@ -34,6 +35,7 @@ Available Commands:
   backup-info     Display information about backups
   completion      Generate the autocompletion script for the specified shell
   help            Help about any command
+  history-clean   Clean failed and deleted backups from the history database
   history-migrate Migrate history database
   report-info     Display the report for a specific backup
 
@@ -55,6 +57,7 @@ Description of each command:
 * [Delete all existing backups older than the specified time condition (`backup-clean`)](./COMMANDS.md#delete-all-existing-backups-older-than-the-specified-time-condition-backup-clean)
 * [Delete a specific existing backup (`backup-delete`)](./COMMANDS.md#delete-a-specific-existing-backup-backup-delete)
 * [Display information about backups (`backup-info`)](./COMMANDS.md#display-information-about-backups-backup-info)
+* [Clean failed and deleted backups from the history database (`history-clean`)](./COMMANDS.md#clean-failed-and-deleted-backups-from-the-history-database-history-clean)
 * [Migrate history database (`history-migrate`)](./COMMANDS.md#migrate-history-database-history-migrate)
 * [Display the report for a specific backup (`report-info`)](./COMMANDS.md#display-the-report-for-a-specific-backup-report-info)
 

diff --git a/cmd/backup_clean.go b/cmd/backup_clean.go
@@ -165,6 +165,15 @@ func cleanBackup() error {
 					}
 					err = backupCleanFilePlugin(backupCleanCascade, beforeTimestamp, backupCleanPluginConfigFile, pluginConfig, &parseHData)
 					if err != nil {
+						// In current implementation, there are cases where some backups were deleted, and some were not.
+						// Foe example, the clean command was executed without --cascade option.
+						// In this case - metadata backup was deleted, but full + incrementals - weren't.
+						// We should update the history file even it error occurred.
+						errUpdateHFile := parseHData.UpdateHistoryFile(hFile)
+						if errUpdateHFile != nil {
+							gplog.Error(textmsg.ErrorTextUnableActionHistoryFile("update", errUpdateHFile))
+						}
+						// It is enough to return only one error.
 						return err
 					}
 				} else {

diff --git a/cmd/backup_delete.go b/cmd/backup_delete.go
@@ -245,7 +245,7 @@ func backupDeleteDBPluginFunc(backupName, pluginConfigPath string, pluginConfig
 	var err error
 	dateDeleted := getCurrentTimestamp()
 	gplog.Info(textmsg.InfoTextBackupDeleteStart(backupName))
-	err = updateDeleteStatus(backupName, gpbckpconfig.DateDeletedInProgress, hDB)
+	err = gpbckpconfig.UpdateDeleteStatus(backupName, gpbckpconfig.DateDeletedInProgress, hDB)
 	if err != nil {
 		gplog.Error(textmsg.ErrorTextUnableSetBackupStatus(gpbckpconfig.DateDeletedInProgress, backupName, err))
 		return err
@@ -256,15 +256,15 @@ func backupDeleteDBPluginFunc(backupName, pluginConfigPath string, pluginConfig
 		gplog.Error(stderr)
 	}
 	if errdel != nil {
-		err = updateDeleteStatus(backupName, gpbckpconfig.DateDeletedPluginFailed, hDB)
+		err = gpbckpconfig.UpdateDeleteStatus(backupName, gpbckpconfig.DateDeletedPluginFailed, hDB)
 		if err != nil {
 			gplog.Error(textmsg.ErrorTextUnableSetBackupStatus(gpbckpconfig.DateDeletedPluginFailed, backupName, err))
 		}
 		gplog.Error(textmsg.ErrorTextUnableDeleteBackup(backupName, errdel))
 		return errdel
 	}
 	gplog.Info(stdout)
-	err = updateDeleteStatus(backupName, dateDeleted, hDB)
+	err = gpbckpconfig.UpdateDeleteStatus(backupName, dateDeleted, hDB)
 	if err != nil {
 		gplog.Error(textmsg.ErrorTextUnableSetBackupStatus(dateDeleted, backupName, err))
 		return err
@@ -384,17 +384,3 @@ func execDeleteBackup(executablePath, deleteBackupPluginCommand, pluginConfigFil
 	err := cmd.Run()
 	return stdout.String(), stderr.String(), err
 }
-
-func updateDeleteStatus(timestamp, deleteStatus string, historyDB *sql.DB) error {
-	tx, _ := historyDB.Begin()
-	_, err := tx.Exec(`UPDATE backups
-		SET date_deleted = ?
-		WHERE timestamp = ?;`,
-		deleteStatus, timestamp)
-	if err != nil {
-		tx.Rollback()
-		return err
-	}
-	err = tx.Commit()
-	return err
-}
diff --git a/cmd/backup_info.go b/cmd/backup_info.go
@@ -26,9 +26,9 @@ var backupInfoCmd = &cobra.Command{
 
 By default, only active backups or backups with deletion status "In progress" from gpbackup_history.db are displayed.
 
-To additional display deleted backups, use the --show-deleted option.
-To additional display failed backups, use the --show-failed option.
-To display all backups, use --show-deleted  and --show-failed options together.
+To additional display deleted backups, use the --deleted option.
+To additional display failed backups, use the --failed option.
+To display all backups, use --deleted and --failed options together.
 
 To display backups of a specific type, use the --type option.
 
@@ -54,13 +54,13 @@ func init() {
 	rootCmd.AddCommand(backupInfoCmd)
 	backupInfoCmd.Flags().BoolVar(
 		&backupInfoShowDeleted,
-		showDeletedFlagName,
+		deletedFlagName,
 		false,
 		"show deleted backups",
 	)
 	backupInfoCmd.Flags().BoolVar(
 		&backupInfoShowFailed,
-		showFailedFlagName,
+		failedFlagName,
 		false,
 		"show failed backups",
 	)

diff --git a/cmd/constants.go b/cmd/constants.go
@@ -26,15 +26,19 @@ const (
 	timestampFlagName            = "timestamp"
 	pluginConfigFileFlagName     = "plugin-config"
 	reportFilePluginPathFlagName = "plugin-report-file-path"
-	showDeletedFlagName          = "show-deleted"
-	showFailedFlagName           = "show-failed"
+	deletedFlagName              = "deleted"
+	failedFlagName               = "failed"
 	cascadeFlagName              = "cascade"
 	forceFlagName                = "force"
 	olderThenDaysFlagName        = "older-than-days"
 	beforeTimestampFlagName      = "before-timestamp"
 	typeFlagName                 = "type"
 
 	exitErrorCode = 1
+
+	// Batch size for deleting from sqlite3.
+	// This is to prevent problem with sqlite3.
+	sqliteDeleteBatchSize = 1000
 )
 
 var (