feat(database): Physical Backup & Restore

press/agent.py

@@ -21,7 +21,11 @@
 
 	from press.press.doctype.agent_job.agent_job import AgentJob
 	from press.press.doctype.app_patch.app_patch import AgentPatchConfig, AppPatch
+	from press.press.doctype.physical_backup_restoration.physical_backup_restoration import (
+		PhysicalBackupRestoration,
+	)
 	from press.press.doctype.site.site import Site
+	from press.press.doctype.site_backup.site_backup import SiteBackup
 
 
 class Agent:
@@ -420,12 +424,65 @@ def archive_site(self, site, site_name=None, force=False):
 			site=site.name,
 		)
 
-	def backup_site(self, site, with_files=False, offsite=False):
+	def physical_backup_database(self, site: Site, site_backup: SiteBackup):
+		"""
+		For physical database backup, the flow :
+		- Create the agent job
+		- Agent job will lock the specific database + flush the changes to disk
+		- Take a database dump
+		- Use `fsync` to ensure the changes are written to disk
+		- Agent will send back a request to FC for taking the snapshot
+			- By calling `snapshot_create_callback` url
+		- Then, unlock the database
+		"""
+		url = frappe.utils.get_url()
+		data = {
+			"databases": [site_backup.database_name],
+			"mariadb_root_password": get_mariadb_root_password(site),
+			"private_ip": frappe.get_value(
+				"Database Server", frappe.db.get_value("Server", site.server, "database_server"), "private_ip"
+			),
+			"site_backup": {
+				"name": site_backup.name,
+				"snapshot_request_key": site_backup.snapshot_request_key,
+				"snapshot_trigger_url": f"{url}/api/method/press.api.site_backup.create_snapshot",
+			},
+		}
+		return self.create_agent_job(
+			"Physical Backup Database",
+			"/database/physical-backup",
+			data=data,
+			bench=site.bench,
+			site=site.name,
+		)
+
+	def physical_restore_database(self, site, backup_restoration: PhysicalBackupRestoration):
+		backup: SiteBackup = frappe.get_doc("Site Backup", backup_restoration.site_backup)
+		data = {
+			"backup_db": backup_restoration.source_database,
+			"target_db": backup_restoration.destination_database,
+			"target_db_root_password": get_mariadb_root_password(site),
+			"innodb_tables": json.loads(backup.innodb_tables),
+			"myisam_tables": json.loads(backup.myisam_tables),
+			"table_schema": backup.table_schema,
+			"backup_db_base_directory": os.path.join(backup_restoration.mount_path, "/var/lib/mysql"),
+		}
+		return self.create_agent_job(
+			"Physical Restore Database",
+			"/database/physical-restore",
+			data=data,
+			bench=site.bench,
+			site=site.name,
+			reference_name=backup_restoration.name,
+			reference_doctype=backup_restoration.doctype,
+		)
+
+	def backup_site(self, site, site_backup: SiteBackup):
 		from press.press.doctype.site_backup.site_backup import get_backup_bucket
 
-		data = {"with_files": with_files}
+		data = {"with_files": site_backup.with_files}
 
-		if offsite:
+		if site_backup.offsite:
 			settings = frappe.get_single("Press Settings")
 			backups_path = os.path.join(site.name, str(date.today()))
 			backup_bucket = get_backup_bucket(site.cluster, region=True)

press/api/site_backup.py

@@ -0,0 +1,31 @@
+from typing import TYPE_CHECKING
+
+import frappe
+
+if TYPE_CHECKING:
+	from press.press.doctype.site_backup.site_backup import SiteBackup
+
+
+@frappe.whitelist(allow_guest=True, methods="POST")
+def create_snapshot(name: str, key: str):
+	"""
+	This API will be called by agent during physical backup of database server.
+	Once, agent prepare the specific database for backup, it will call this API to create a snapshot of the database.
+	Because we need to hold the lock on the database for the duration of the backup.
+	Only after the snapshot is created, the agent will release the lock on the database.
+	"""
+	current_user = frappe.session.user
+	try:
+		frappe.set_user("Administrator")
+		site_backup: SiteBackup = frappe.get_doc("Site Backup", name)
+		if not (key and site_backup.snapshot_request_key == key):
+			frappe.throw("Invalid key for snapshot creation")
+		site_backup.create_database_snapshot()
+		site_backup.reload()
+		# Re-verify if the snapshot was created and linked to the site backup
+		if not site_backup.database_snapshot:
+			frappe.throw("Failed to create a snapshot for the database server")
+	except Exception as e:
+		raise e
+	finally:
+		frappe.set_user(current_user)

press/fixtures/agent_job_type.json

@@ -2290,5 +2290,35 @@
      "step_name": "Analyze Slow Queries"
     }
   ]
+ },
+ {
+  "disabled_auto_retry": 1,
+  "docstatus": 0,
+  "doctype": "Agent Job Type",
+  "max_retry_count": 3,
+  "modified": "2025-01-09 17:07:45.359754",
+  "name": "Physical Backup Database",
+  "request_method": "POST",
+  "request_path": "/database/physical-backup",
+  "steps": [
+   {
+    "step_name": "Fetch Database Tables Information"
+   },
+   {
+    "step_name": "Flush Database Tables"
+   },
+   {
+    "step_name": "Flush Changes to Disk"
+   },
+   {
+    "step_name": "Export Table Schema"
+   },
+   {
+    "step_name": "Create Database Snapshot"
+   },
+   {
+    "step_name": "Unlock Tables"
+   }
+  ]
  }
 ]

press/press/doctype/agent_job/agent_job.py

@@ -959,6 +959,9 @@ def process_job_updates(job_name: str, response_data: dict | None = None):  # no
 			process_stop_code_server_job_update,
 		)
 		from press.press.doctype.deploy_candidate.deploy_candidate import DeployCandidate
+		from press.press.doctype.physical_backup_restoration.physical_backup_restoration import (
+			process_job_update as process_physical_backup_restoration_job_update,
+		)
 		from press.press.doctype.proxy_server.proxy_server import (
 			process_update_nginx_job_update,
 		)
@@ -1022,7 +1025,7 @@ def process_job_updates(job_name: str, response_data: dict | None = None):  # no
 			process_stop_code_server_job_update(job)
 		elif job.job_type == "Archive Code Server" or job.job_type == "Remove Code Server from Upstream":
 			process_archive_code_server_job_update(job)
-		elif job.job_type == "Backup Site":
+		elif job.job_type in ["Backup Site", "Physical Backup Database"]:
 			process_backup_site_job_update(job)
 		elif job.job_type == "Archive Site" or job.job_type == "Remove Site from Upstream":
 			process_archive_site_job_update(job)
@@ -1073,6 +1076,8 @@ def process_job_updates(job_name: str, response_data: dict | None = None):  # no
 			"Modify Database User Permissions",
 		]:
 			SiteDatabaseUser.process_job_update(job)
+		elif job.job_type == "Physical Restore Database":
+			process_physical_backup_restoration_job_update(job)
 
 		# send failure notification if job failed
 		if job.status == "Failure":

press/press/doctype/physical_backup_restoration/physical_backup_restoration.js

@@ -0,0 +1,8 @@
+// Copyright (c) 2025, Frappe and contributors
+// For license information, please see license.txt
+
+// frappe.ui.form.on("Physical Backup Restoration", {
+// 	refresh(frm) {
+
+// 	},
+// });

press/press/doctype/physical_backup_restoration/physical_backup_restoration.json

@@ -0,0 +1,146 @@
+{
+ "actions": [],
+ "allow_rename": 1,
+ "creation": "2025-01-10 13:02:39.393157",
+ "doctype": "DocType",
+ "engine": "InnoDB",
+ "field_order": [
+  "site",
+  "status",
+  "site_backup",
+  "disk_snapshot",
+  "column_break_zind",
+  "volume",
+  "device",
+  "mount_path",
+  "job",
+  "section_break_pqgo",
+  "source_database",
+  "column_break_kaja",
+  "destination_database",
+  "destination_server",
+  "section_break_aqam",
+  "steps"
+ ],
+ "fields": [
+  {
+   "fieldname": "site",
+   "fieldtype": "Link",
+   "in_list_view": 1,
+   "label": "Site",
+   "options": "Site",
+   "reqd": 1
+  },
+  {
+   "fieldname": "column_break_zind",
+   "fieldtype": "Column Break"
+  },
+  {
+   "default": "Pending",
+   "fieldname": "status",
+   "fieldtype": "Select",
+   "in_list_view": 1,
+   "label": "Status",
+   "options": "Pending\nRunning\nSuccess\nFailure",
+   "reqd": 1
+  },
+  {
+   "fieldname": "disk_snapshot",
+   "fieldtype": "Link",
+   "label": "Disk Snapshot",
+   "options": "Virtual Disk Snapshot",
+   "reqd": 1
+  },
+  {
+   "fieldname": "site_backup",
+   "fieldtype": "Link",
+   "label": "Site Backup",
+   "options": "Site Backup"
+  },
+  {
+   "fieldname": "section_break_pqgo",
+   "fieldtype": "Section Break"
+  },
+  {
+   "fieldname": "source_database",
+   "fieldtype": "Data",
+   "label": "Source Database",
+   "reqd": 1
+  },
+  {
+   "fieldname": "column_break_kaja",
+   "fieldtype": "Column Break"
+  },
+  {
+   "fieldname": "destination_database",
+   "fieldtype": "Data",
+   "label": " Destination Database",
+   "reqd": 1
+  },
+  {
+   "fieldname": "section_break_aqam",
+   "fieldtype": "Section Break"
+  },
+  {
+   "fieldname": "steps",
+   "fieldtype": "Table",
+   "label": "Steps",
+   "options": "Physical Backup Restoration Step"
+  },
+  {
+   "fieldname": "destination_server",
+   "fieldtype": "Link",
+   "label": "Destination Server",
+   "options": "Database Server",
+   "reqd": 1
+  },
+  {
+   "fieldname": "volume",
+   "fieldtype": "Data",
+   "label": "Volume",
+   "read_only": 1
+  },
+  {
+   "fieldname": "device",
+   "fieldtype": "Data",
+   "label": "Device",
+   "read_only": 1
+  },
+  {
+   "fieldname": "mount_path",
+   "fieldtype": "Data",
+   "label": "Mount Path",
+   "read_only": 1
+  },
+  {
+   "fieldname": "job",
+   "fieldtype": "Link",
+   "label": "Job",
+   "options": "Agent Job"
+  }
+ ],
+ "index_web_pages_for_search": 1,
+ "links": [],
+ "modified": "2025-01-17 19:24:16.834232",
+ "modified_by": "Administrator",
+ "module": "Press",
+ "name": "Physical Backup Restoration",
+ "owner": "Administrator",
+ "permissions": [
+  {
+   "create": 1,
+   "delete": 1,
+   "email": 1,
+   "export": 1,
+   "print": 1,
+   "read": 1,
+   "report": 1,
+   "role": "System Manager",
+   "share": 1,
+   "write": 1
+  }
+ ],
+ "sort_field": "creation",
+ "sort_order": "DESC",
+ "states": []
+}