16 Commits

Author SHA1 Message Date
9e364af93e Update README.md 2026-01-14 15:06:28 +01:00
d88c572b56 Update README.md 2026-01-14 12:11:18 +01:00
d8a9f80df6 Update README.md 2026-01-14 11:49:20 +01:00
cac84051e8 Update backup.py 2026-01-14 11:44:10 +01:00
e981912836 Remove legacy put_stream method
Removed legacy chunked upload method 'put_stream' in favor of 'put_file' for better proxy compatibility.
2026-01-14 11:43:38 +01:00
e5b7f9c373 Update version in manifest.json to 0.2.0 2026-01-14 11:42:48 +01:00
88e7bb8d47 Update project name to 'Owncloud Backup Homeassistant' 2026-01-14 11:42:24 +01:00
ca49174419 Revise README for project title and version update
Updated project title, version status, and added sections on upload reliability and troubleshooting.
2026-01-14 11:42:03 +01:00
53473f1f0f Update CHANGELOG for version 0.2.0
Added new features and fixes for version 0.2.0, including improved compatibility and upload reliability.
2026-01-14 11:40:56 +01:00
0022558d25 Update issue tracker URL in manifest.json
Updated the issue tracker URL in the manifest file.
2026-01-14 10:50:01 +01:00
b59d84bacc Update manifest.json 2026-01-14 10:49:27 +01:00
4171159264 Implement long operation timeouts for WebDAV client
Added a timeout for long WebDAV operations to improve reliability.
2026-01-14 10:44:49 +01:00
a67a631c99 Implement spooling of byte stream to temp file
Added functionality to spool an async byte stream to a temporary file for improved upload handling.
2026-01-14 10:44:24 +01:00
99c362b6d4 Add SPOOL_FLUSH_BYTES constant for uploads
Added constant for spooling to temporary file during uploads.
2026-01-14 10:44:00 +01:00
76715585ab Update manifest.json for version 0.1.1-alpha
Updated version number and links in manifest.
2026-01-14 10:43:18 +01:00
15e6ae9ab7 Update CHANGELOG for version 0.1.1-alpha
Improved upload reliability and adjusted client timeout for WebDAV.
2026-01-14 10:42:24 +01:00
6 changed files with 253 additions and 56 deletions

View File

@@ -4,6 +4,21 @@ All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
## [0.2.0] - 2026-01-14
### Added
- Improved cross-version compatibility with Home Assistant backup metadata by normalizing backup schema fields (e.g., `addons`, `database_included`, etc.).
- More robust metadata serialization for `AgentBackup` across Home Assistant versions (supports different serialization methods).
### Fixed
- Improved upload reliability by spooling backup streams to a temporary file and uploading with Content-Length (avoids chunked WebDAV uploads that may cause reverse proxy 504 timeouts).
- Added non-restrictive client timeouts for long-running WebDAV operations to prevent client-side aborts.
- Fixed backup listing failures caused by missing expected metadata keys in different Home Assistant versions.
## [0.1.1-alpha] - 2026-01-14
### Fixed
- Improved upload reliability by spooling backup streams to a temporary file and uploading with Content-Length (avoids chunked WebDAV uploads that may cause reverse proxy 504 timeouts).
- Set a non-restrictive client timeout for WebDAV PUT requests to prevent client-side premature aborts on slow connections.
## [0.1.0-alpha] - 2026-01-14
### Added
- Initial alpha release

View File

@@ -1,4 +1,10 @@
# Owncloud Backup Homeassistant by René Bachmann
# Owncloud Backup Homeassistant
[![Latest Release](https://img.shields.io/github/v/release/bahmcloud/owncloud-backup-ha?style=flat-square)](https://github.com/bahmcloud/owncloud-backup-ha/releases)
[![HACS](https://img.shields.io/badge/HACS-Custom-orange.svg?style=flat-square)](https://hacs.xyz/)
[![License](https://img.shields.io/github/license/bahmcloud/owncloud-backup-ha?style=flat-square)](LICENSE)
[![Open in HACS](https://my.home-assistant.io/badges/hacs_repository.svg)](https://my.home-assistant.io/redirect/hacs_repository/?owner=bahmcloud&repository=owncloud-backup-ha&category=integration)
Home Assistant custom integration that adds **ownCloud (Classic/Server)** as a **Backup Location / Backup Agent** using the official **WebDAV** interface.
@@ -8,8 +14,8 @@ This integration allows you to:
- download and **restore** backups via the Home Assistant UI
- authenticate using **either** an ownCloud **App Password** (recommended for 2FA) **or** the regular account password
> **Status:** `0.1.0-alpha`
> This is an early alpha release. Please test on a non-critical system first.
> **Status:** `0.2.0`
> This release focuses on reliability and compatibility across Home Assistant versions.
---
@@ -28,6 +34,15 @@ This integration allows you to:
- ✅ English UI & documentation
- ✅ HACS-ready repository structure
### Upload reliability (important)
To improve reliability behind reverse proxies and avoid WebDAV timeouts with chunked uploads,
the integration **spools the backup to a temporary file** and then uploads it with a proper
**Content-Length** header.
### Home Assistant compatibility
Home Assistant has evolved its backup metadata schema over time. This integration normalizes
backup metadata keys to remain compatible across multiple Home Assistant versions.
---
## Requirements
@@ -41,12 +56,13 @@ This integration allows you to:
### Add as a custom repository
1. In Home Assistant: **HACS → Integrations → ⋮ → Custom repositories**
2. Add your repository URL
3. Category: **Integration**
4. Install **owncloud-backup-ha**
5. Restart Home Assistant
> This repository includes a `hacs.json` and follows the required `custom_components/` layout.
2. Add repository URL:
```
https://github.com/bahmcloud/owncloud-backup-ha
```
4. Category: **Integration**
5. Install **owncloud-backup-ha**
6. Restart Home Assistant
---
@@ -101,6 +117,15 @@ Home Assistant will download the `.tar` from ownCloud using the Backup Agent API
## Troubleshooting
### "Upload failed" / HTTP 504 (Gateway Timeout)
A 504 typically indicates a reverse proxy / gateway timeout (e.g., Nginx/Traefik/Cloudflare).
This integration uploads with Content-Length (non-chunked) for better compatibility.
If you still see 504:
- Increase proxy timeouts (e.g. `proxy_read_timeout`, `proxy_send_timeout` in Nginx)
- Ensure large uploads are allowed (`client_max_body_size` in Nginx)
- Avoid buffering restrictions for WebDAV endpoints
### "Cannot connect"
- Check your **Base URL**
- Make sure the ownCloud user can access WebDAV
@@ -112,12 +137,6 @@ Home Assistant will download the `.tar` from ownCloud using the Backup Agent API
- install the CA properly, or
- temporarily disable **Verify SSL** (not recommended for production)
### Missing backups in list
- Ensure the configured backup folder is correct
- Check that ownCloud contains either:
- `.json` metadata files (preferred), or
- `.tar` files (fallback)
---
## Security notes

View File

@@ -3,7 +3,10 @@ from __future__ import annotations
import asyncio
import json
import logging
import os
import tempfile
from collections.abc import AsyncIterator, Callable, Coroutine
from dataclasses import asdict, is_dataclass
from typing import Any
from homeassistant.components.backup import (
@@ -12,7 +15,6 @@ from homeassistant.components.backup import (
BackupAgentError,
BackupNotFound,
)
from homeassistant.core import HomeAssistant, callback
from .const import (
@@ -20,6 +22,7 @@ from .const import (
DATA_CLIENT,
DOMAIN,
META_SUFFIX,
SPOOL_FLUSH_BYTES,
TAR_PREFIX,
TAR_SUFFIX,
)
@@ -36,14 +39,112 @@ def _make_meta_name(backup_id: str) -> str:
return f"{TAR_PREFIX}{backup_id}{META_SUFFIX}"
def _normalize_backup_dict(d: dict[str, Any]) -> dict[str, Any]:
"""Normalize backup metadata to satisfy multiple HA schema versions."""
d = dict(d)
# Identity fields (varies by HA versions)
d.setdefault("backup_id", d.get("slug", ""))
d.setdefault("slug", d.get("backup_id", ""))
# Presentation fields
d.setdefault("name", f"ownCloud backup ({d.get('backup_id') or d.get('slug') or 'unknown'})")
d.setdefault("date", d.get("created_at", ""))
d.setdefault("size", 0)
d.setdefault("protected", False)
d.setdefault("compressed", True)
d.setdefault("extra_metadata", {})
# Content selections
d.setdefault("addons", [])
d.setdefault("folders", [])
# Older schema booleans
d.setdefault("database", True)
d.setdefault("homeassistant", True)
# Newer schema booleans
d.setdefault("database_included", d.get("database", True))
d.setdefault("homeassistant_included", d.get("homeassistant", True))
d.setdefault("addons_included", bool(d.get("addons", [])))
d.setdefault("folders_included", bool(d.get("folders", [])))
# Keep old keys consistent
d["database"] = bool(d.get("database_included", True))
d["homeassistant"] = bool(d.get("homeassistant_included", True))
if not isinstance(d.get("addons"), list):
d["addons"] = []
if not isinstance(d.get("folders"), list):
d["folders"] = []
return d
def _agentbackup_to_dict(backup: AgentBackup) -> dict[str, Any]:
"""Serialize AgentBackup in a HA-version-independent way."""
if hasattr(backup, "to_dict"):
raw = backup.to_dict() # type: ignore[assignment]
elif hasattr(backup, "as_dict"):
raw = backup.as_dict() # type: ignore[assignment]
elif is_dataclass(backup):
raw = asdict(backup)
else:
raw = {k: v for k, v in vars(backup).items() if not k.startswith("_")}
return _normalize_backup_dict(raw)
def _agentbackup_from_dict(d: dict[str, Any]) -> AgentBackup:
"""Best-effort create AgentBackup across HA versions."""
"""Create AgentBackup from dict and ensure required keys exist."""
d = _normalize_backup_dict(d)
from_dict = getattr(AgentBackup, "from_dict", None)
if callable(from_dict):
return from_dict(d) # type: ignore[misc]
return AgentBackup(**d) # type: ignore[arg-type]
async def _spool_stream_to_tempfile(stream: AsyncIterator[bytes]) -> tuple[str, int]:
"""Spool an async byte stream into a temporary file and return (path, size)."""
fd, path = tempfile.mkstemp(prefix="owncloud_backup_", suffix=".tar")
os.close(fd)
size = 0
buf = bytearray()
try:
async for chunk in stream:
if not chunk:
continue
buf.extend(chunk)
size += len(chunk)
if len(buf) >= SPOOL_FLUSH_BYTES:
data = bytes(buf)
buf.clear()
await asyncio.to_thread(_write_bytes_to_file, path, data, append=True)
if buf:
await asyncio.to_thread(_write_bytes_to_file, path, bytes(buf), append=True)
return path, size
except Exception:
try:
os.remove(path)
except OSError:
pass
raise
def _write_bytes_to_file(path: str, data: bytes, *, append: bool) -> None:
mode = "ab" if append else "wb"
with open(path, mode) as f:
f.write(data)
f.flush()
class OwnCloudBackupAgent(BackupAgent):
"""Backup agent storing backups in ownCloud via WebDAV."""
@@ -61,21 +162,32 @@ class OwnCloudBackupAgent(BackupAgent):
backup: AgentBackup,
**kwargs: Any,
) -> None:
"""Upload a backup + metadata sidecar."""
"""Upload a backup + metadata sidecar (spooled for non-chunked PUT)."""
temp_path: str | None = None
try:
tar_name = _make_tar_name(backup.backup_id)
meta_name = _make_meta_name(backup.backup_id)
# 1) Upload tar stream
# 1) Spool tar stream to temp file
stream = await open_stream()
await self._client.put_stream(tar_name, stream)
temp_path, size = await _spool_stream_to_tempfile(stream)
# 2) Upload metadata JSON (small)
meta_bytes = json.dumps(backup.to_dict(), ensure_ascii=False).encode("utf-8")
# 2) Upload tar file with Content-Length
await self._client.put_file(tar_name, temp_path, size)
# 3) Upload normalized metadata JSON
meta_dict = _agentbackup_to_dict(backup)
meta_bytes = json.dumps(meta_dict, ensure_ascii=False).encode("utf-8")
await self._client.put_bytes(meta_name, meta_bytes)
except Exception as err: # noqa: BLE001
raise BackupAgentError(f"Upload to ownCloud failed: {err}") from err
finally:
if temp_path:
try:
os.remove(temp_path)
except OSError:
pass
async def async_list_backups(self, **kwargs: Any) -> list[AgentBackup]:
"""List backups by reading metadata sidecars; fallback to tar stat if missing."""
@@ -87,7 +199,6 @@ class OwnCloudBackupAgent(BackupAgent):
backups: list[AgentBackup] = []
# 1) Load metadata sidecars (limited concurrency)
sem = asyncio.Semaphore(5)
async def fetch_meta(meta_name: str) -> None:
@@ -101,7 +212,6 @@ class OwnCloudBackupAgent(BackupAgent):
await asyncio.gather(*(fetch_meta(m) for m in meta_files))
# 2) Fallback: tar without meta -> synthesize minimal AgentBackup
known_ids = {b.backup_id for b in backups}
for tar_name in tar_files:
backup_id = tar_name.removeprefix(TAR_PREFIX).removesuffix(TAR_SUFFIX)
@@ -109,13 +219,21 @@ class OwnCloudBackupAgent(BackupAgent):
continue
info = await self._client.stat(tar_name)
d = {
d = _normalize_backup_dict(
{
"backup_id": backup_id,
"name": f"ownCloud backup ({backup_id})",
"date": info.get("modified_iso", ""),
"size": info.get("size", 0),
"protected": False,
# conservative defaults:
"database_included": True,
"homeassistant_included": True,
"addons_included": False,
"folders_included": False,
"addons": [],
"folders": [],
}
)
backups.append(_agentbackup_from_dict(d))
backups.sort(key=lambda b: str(b.date), reverse=True)
@@ -129,7 +247,6 @@ class OwnCloudBackupAgent(BackupAgent):
meta_name = _make_meta_name(backup_id)
tar_name = _make_tar_name(backup_id)
# 1) Try meta
try:
raw = await self._client.get_bytes(meta_name)
d = json.loads(raw.decode("utf-8"))
@@ -139,16 +256,22 @@ class OwnCloudBackupAgent(BackupAgent):
except Exception as err: # noqa: BLE001
raise BackupAgentError(f"Get backup metadata failed: {err}") from err
# 2) Fallback to tar stat
try:
info = await self._client.stat(tar_name)
d = {
d = _normalize_backup_dict(
{
"backup_id": backup_id,
"name": f"ownCloud backup ({backup_id})",
"date": info.get("modified_iso", ""),
"size": info.get("size", 0),
"protected": False,
"database_included": True,
"homeassistant_included": True,
"addons_included": False,
"folders_included": False,
"addons": [],
"folders": [],
}
)
return _agentbackup_from_dict(d)
except FileNotFoundError as err:
raise BackupNotFound(f"Backup not found: {backup_id}") from err

View File

@@ -14,3 +14,6 @@ DATA_BACKUP_AGENT_LISTENERS = "backup_agent_listeners"
TAR_PREFIX = "ha_backup_"
TAR_SUFFIX = ".tar"
META_SUFFIX = ".json"
# Spooling to temp file to avoid chunked WebDAV uploads
SPOOL_FLUSH_BYTES = 1024 * 1024 # 1 MiB

View File

@@ -1,10 +1,10 @@
{
"domain": "owncloud_backup",
"name": "ownCloud Backup (WebDAV)",
"version": "0.1.0-alpha",
"documentation": "https://github.com/your-org/owncloud-backup-ha",
"issue_tracker": "https://github.com/your-org/owncloud-backup-ha/issues",
"codeowners": [],
"version": "0.2.0",
"documentation": "https://github.com/bahmcloud/owncloud-backup-ha/",
"issue_tracker": "https://github.com/bahmcloud/owncloud-backup-ha/issues",
"codeowners": ["@bahmcloud"],
"config_flow": true,
"integration_type": "service",
"iot_class": "cloud_push",

View File

@@ -2,6 +2,7 @@ from __future__ import annotations
import base64
import logging
import os
import xml.etree.ElementTree as ET
from collections.abc import AsyncIterator
from datetime import datetime, timezone
@@ -9,6 +10,7 @@ from email.utils import parsedate_to_datetime
from typing import Final
from urllib.parse import quote, urljoin
import aiohttp
from aiohttp import ClientResponseError, ClientSession
from yarl import URL
@@ -46,6 +48,11 @@ class WebDavClient:
]
self._cached_root: str | None = None
# Non-restrictive client timeouts for potentially long WebDAV operations
self._timeout_long = aiohttp.ClientTimeout(
total=None, connect=60, sock_connect=60, sock_read=None
)
def _auth_header(self) -> str:
token = base64.b64encode(f"{self._username}:{self._password}".encode("utf-8")).decode("ascii")
return f"Basic {token}"
@@ -82,6 +89,7 @@ class WebDavClient:
b'<d:propfind xmlns:d="DAV:"><d:prop><d:resourcetype/></d:prop></d:propfind>'
),
raise_for_status=True,
timeout=self._timeout_long,
):
self._cached_root = root
return root
@@ -106,6 +114,7 @@ class WebDavClient:
base_folder,
headers=self._headers({"Depth": "0"}),
raise_for_status=True,
timeout=self._timeout_long,
):
return
except ClientResponseError as err:
@@ -127,7 +136,11 @@ class WebDavClient:
# exists?
try:
async with self._session.request(
"PROPFIND", url, headers=self._headers({"Depth": "0"}), raise_for_status=True
"PROPFIND",
url,
headers=self._headers({"Depth": "0"}),
raise_for_status=True,
timeout=self._timeout_long,
):
return
except ClientResponseError as err:
@@ -135,7 +148,9 @@ class WebDavClient:
raise
# create
async with self._session.request("MKCOL", url, headers=self._headers()) as resp:
async with self._session.request(
"MKCOL", url, headers=self._headers(), timeout=self._timeout_long
) as resp:
if resp.status in (201, 405):
return
text = await resp.text()
@@ -154,6 +169,7 @@ class WebDavClient:
b'<d:propfind xmlns:d="DAV:"><d:prop><d:displayname/></d:prop></d:propfind>'
),
raise_for_status=True,
timeout=self._timeout_long,
) as resp:
body = await resp.text()
@@ -195,24 +211,44 @@ class WebDavClient:
async def put_bytes(self, name: str, data: bytes) -> None:
folder = await self._base_folder_url()
url = self._file_url(folder, name)
async with self._session.put(url, data=data, headers=self._headers(), raise_for_status=True):
async with self._session.put(
url,
data=data,
headers=self._headers({"Content-Length": str(len(data))}),
raise_for_status=True,
timeout=self._timeout_long,
):
return
async def put_stream(self, name: str, stream: AsyncIterator[bytes]) -> None:
async def put_file(self, name: str, path: str, size: int) -> None:
"""Upload a local file with an explicit Content-Length (non-chunked)."""
folder = await self._base_folder_url()
url = self._file_url(folder, name)
async def gen():
async for chunk in stream:
yield chunk
# Ensure correct size if caller passes 0/unknown
if size <= 0:
try:
size = os.path.getsize(path)
except OSError:
size = 0
async with self._session.put(url, data=gen(), headers=self._headers(), raise_for_status=True):
headers = {"Content-Length": str(size)} if size > 0 else {}
# aiohttp will stream file content; with Content-Length set, proxies are usually happier.
with open(path, "rb") as f:
async with self._session.put(
url,
data=f,
headers=self._headers(headers),
raise_for_status=True,
timeout=self._timeout_long,
):
return
async def get_bytes(self, name: str) -> bytes:
folder = await self._base_folder_url()
url = self._file_url(folder, name)
async with self._session.get(url, headers=self._headers()) as resp:
async with self._session.get(url, headers=self._headers(), timeout=self._timeout_long) as resp:
if resp.status == 404:
raise FileNotFoundError(name)
resp.raise_for_status()
@@ -221,7 +257,7 @@ class WebDavClient:
async def get_stream(self, name: str) -> AsyncIterator[bytes]:
folder = await self._base_folder_url()
url = self._file_url(folder, name)
resp = await self._session.get(url, headers=self._headers())
resp = await self._session.get(url, headers=self._headers(), timeout=self._timeout_long)
if resp.status == 404:
await resp.release()
raise FileNotFoundError(name)
@@ -239,7 +275,7 @@ class WebDavClient:
async def delete(self, name: str) -> None:
folder = await self._base_folder_url()
url = self._file_url(folder, name)
async with self._session.delete(url, headers=self._headers()) as resp:
async with self._session.delete(url, headers=self._headers(), timeout=self._timeout_long) as resp:
if resp.status == 404:
raise FileNotFoundError(name)
if resp.status in (200, 202, 204):
@@ -267,6 +303,7 @@ class WebDavClient:
url,
headers=self._headers({"Depth": "0", "Content-Type": "application/xml; charset=utf-8"}),
data=body,
timeout=self._timeout_long,
) as resp:
if resp.status == 404:
raise FileNotFoundError(name)