less file shares

This commit is contained in:
Ludwig Lehnert
2026-03-17 09:53:56 +01:00
parent 9bf0694bf2
commit 029488b80d
9 changed files with 561 additions and 231 deletions

View File

@@ -5,7 +5,6 @@ JOIN_USER=FileShare_ServiceAcc
JOIN_PASSWORD=ReplaceWithLongRandomPassword
DOMAIN_USERS_SID=S-1-5-21-1111111111-2222222222-3333333333-513
DOMAIN_ADMINS_SID=S-1-5-21-1111111111-2222222222-3333333333-512
PUBLIC_GROUP_SID=S-1-5-21-1111111111-2222222222-3333333333-513
FSLOGIX_GROUP_SID=S-1-5-21-1111111111-2222222222-3333333333-513
# SAMBA_HOSTNAME=adsambafsrv
# NETBIOS_NAME=ADSAMBAFSRV
@@ -13,3 +12,7 @@ FSLOGIX_GROUP_SID=S-1-5-21-1111111111-2222222222-3333333333-513
# LDAP_BASE_DN=DC=example,DC=com
# PRIVATE_SKIP_USERS=svc_backup,svc_sql
# PRIVATE_SKIP_PREFIXES=svc_,sql_
# BACKUP_DESTINATION=rsync://user:pass@backup.example.com/samba-backups
# BACKUP_DESTINATION=smb://DOMAIN%5Cuser:pass@backup.example.com/Backups/samba
# BACKUP_DESTINATION=davfs://user:pass@webdav.example.com/remote.php/dav/files/backup
# BACKUP_DESTINATION=sftp://user:pass@sftp.example.com/exports/samba

View File

@@ -12,6 +12,8 @@ RUN apt-get update \
libnss-winbind \
libpam-winbind \
python3 \
rclone \
rsync \
samba \
samba-vfs-modules \
sqlite3 \
@@ -19,14 +21,15 @@ RUN apt-get update \
winbind \
&& rm -rf /var/lib/apt/lists/*
RUN mkdir -p /app /data/private /data/public /data/fslogix /data/groups /state /etc/samba/generated
RUN mkdir -p /app /data/private /data/fslogix /data/groups/data /data/groups/archive /state
COPY app/reconcile_shares.py /app/reconcile_shares.py
COPY app/backup_to_destination.py /app/backup_to_destination.py
COPY app/init.sh /app/init.sh
COPY etc/samba/smb.conf /app/smb.conf.template
RUN chmod +x /app/init.sh /app/reconcile_shares.py \
&& touch /etc/samba/generated/shares.conf
RUN chmod +x /app/init.sh /app/reconcile_shares.py /app/backup_to_destination.py \
&& true
ENTRYPOINT ["/usr/bin/tini", "--"]
CMD ["/app/init.sh"]

View File

@@ -1,18 +1,19 @@
# AD-Integrated Containerized Samba File Server
This repository provides a production-oriented Samba file server container that joins an existing Active Directory domain, exposes static and dynamic SMB shares, and persists share identity by AD `objectGUID`.
This repository provides a production-oriented Samba file server container that joins an existing Active Directory domain and exposes three SMB shares: `Privat`, `Data`, and `FSLogix`.
## Architecture
- Samba runs in ADS mode with `winbind` identity mapping.
- Static shares:
- `\\server\Privat` -> `/data/private`
- `\\server\Geteilt` -> `/data/public`
- `\\server\Data` -> `/data/groups/data`
- `\\server\FSLogix` -> `/data/fslogix`
- Dynamic shares are generated from AD groups matching `FileShare_*` or `FS_*` and written to `/etc/samba/generated/shares.conf`.
- Dynamic share records are persisted in SQLite at `/state/shares.db`.
- Backing storage is GUID-based and stable across group rename:
- `/data/groups/<objectGUID>`
- FS_* groups are projected as folders inside the Data share (`/data/groups/data/<groupName>`).
- Group records are persisted in SQLite at `/state/shares.db`.
- Group folders are name-based while active and moved to archive on deactivation:
- active: `/data/groups/data/<groupName>`
- inactive/deleted groups: `/data/groups/archive/<groupName>`
- Samba machine trust/key material is persisted in `/var/lib/samba` to survive container recreation.
- Container hostname is fixed (`SAMBA_HOSTNAME`) to keep AD computer identity stable.
- NetBIOS name defaults to `ADSAMBAFSRV` and is clamped to 15 characters (`NETBIOS_NAME` override supported).
@@ -20,19 +21,23 @@ This repository provides a production-oriented Samba file server container that
- `FSLOGIX_GROUP_SID` controls who can access the default FSLogix share (defaults to `DOMAIN_USERS_SID`).
- Startup resolves those SIDs to NSS group names via winbind, then uses those resolved groups in Samba `valid users` rules.
- Share operations are audited with Samba `full_audit` (connect, list, read, write, create, delete, rename) and written to Samba log files.
- Optional remote backups run when `BACKUP_DESTINATION` is configured.
- Private home creation skips well-known/service accounts by default (including `krbtgt`, `msol_*`, `FileShare_ServiceAcc`).
- Reconciliation is executed:
- once on startup
- every 5 minutes via cron
- Backup is executed:
- once on startup (when enabled)
- every 30 minutes via cron
## Dynamic Share Lifecycle (objectGUID-first)
## Data Folder Lifecycle
The reconciliation script (`/app/reconcile_shares.py`) enforces these rules:
1. New matching group -> insert DB row, create `/data/groups/<objectGUID>`, expose share.
2. Group renamed but still `FileShare_` -> update `samAccountName` and `shareName`, keep same path.
3. Group removed or no longer matches prefix -> set `isActive=0`, remove Samba exposure, keep data.
4. Previously inactive group returns -> set `isActive=1`, reuse existing path and data.
1. New matching `FS_*` group -> insert DB row and create `/data/groups/data/<groupName>`.
2. Group rename while still matching `FS_*` -> rename/update folder path.
3. Group removed or no longer matching `FS_*` -> set `isActive=0` and move folder to `/data/groups/archive/...`.
4. Previously inactive group returns -> set `isActive=1`, move back into `/data/groups/data/...`.
## SQLite State Database
@@ -58,8 +63,8 @@ CREATE TABLE shares (
- Initial admin credentials with rights to create/reset `FileShare_ServiceAccount` during `./setup`.
- `FileShare_ServiceAccount` must be allowed to join computers to the domain (`net ads join`) in your AD policy.
- Dynamic group discovery primarily uses machine-account LDAP (`net ads search -P`); join credentials are only used as a fallback LDAP bind path.
- Group naming convention for dynamic shares:
- `FileShare_<ShareName>` or `FS_<ShareName>`
- Group naming convention for Data folders:
- `FS_<FolderName>`
## DNS Requirements
@@ -84,8 +89,8 @@ Kerberos requires close time alignment.
- `README.md`
- `app/init.sh`
- `app/reconcile_shares.py`
- `app/backup_to_destination.py`
- `etc/samba/smb.conf`
- `etc/samba/generated/`
## Setup
@@ -102,8 +107,8 @@ Kerberos requires close time alignment.
- initial admin credentials (used once for provisioning)
- `DOMAIN_USERS_SID`
- `DOMAIN_ADMINS_SID`
- optional `PUBLIC_GROUP_SID` (defaults to `DOMAIN_USERS_SID`)
- optional `FSLOGIX_GROUP_SID` (defaults to `DOMAIN_USERS_SID`)
- optional `BACKUP_DESTINATION` (empty disables backup)
Optional:
- `SAMBA_HOSTNAME` (defaults to `adsambafsrv`)
@@ -144,31 +149,43 @@ Kerberos requires close time alignment.
- mode: `700`
- `hide unreadable = yes` + ACLs enforce that users only see their own folder.
### Geteilt
### Data
- Share: `\\server\Geteilt`
- Path: `/data/public`
- Read/write for authenticated users in configurable `PUBLIC_GROUP_SID` (default: `DOMAIN_USERS_SID`, resolved through winbind).
- Share: `\\server\Data`
- Path: `/data/groups/data`
- Contains one folder per active `FS_*` AD group.
- Root is discoverable as one share, while access to each group folder is enforced via POSIX/ACL group permissions.
- No guest access.
- Permissions are reconciled recursively so all descendants remain homogeneous (dirs `2770`, files `0660`, shared group/admin ACLs).
### FSLogix
- Share: `\\server\FSLogix`
- Path: `/data/fslogix`
- Access for authenticated users in configurable `FSLOGIX_GROUP_SID` (default: `DOMAIN_USERS_SID`, resolved through winbind).
- Semantics intentionally differ from `Geteilt`: only the share root is reconciled (`03770` + ACL defaults), while user-created profile container folders/files are not recursively normalized.
- Semantics intentionally differ from `Data`: only the share root is reconciled (`03770` + ACL defaults), while user-created profile container folders/files are not recursively normalized.
- Samba masks are profile-container oriented (`create mask = 0600`, `directory mask = 0700`) so profile payload stays user-private by default.
### Dynamic Group Shares
## Backups
- AD groups: `FileShare_*` and `FS_*`
- Share name: group title from AD (`displayName` -> `name`/`cn` fallback). Prefix stripping from `sAMAccountName` is only a fallback when no title exists.
- Backing path: `/data/groups/<objectGUID>`
- Share exposure generated in `/etc/samba/generated/shares.conf`
- Dynamic share names are validated for SMB compatibility and deduplicated case-insensitively.
- Group membership changes are refreshed continuously via winbind cache updates (`winbind cache time = 60`) and Samba config reload during reconciliation.
- Dynamic share trees are reconciled recursively so all descendants keep homogeneous permissions.
- Backups are enabled only if `BACKUP_DESTINATION` is non-empty.
- Sources synced to destination on each run:
- `/data/private` -> `data/private`
- `/data/groups` -> `data/groups`
- `/data/fslogix` -> `data/fslogix`
- `/state` -> `state`
- `/var/lib/samba/private` -> `samba/private`
- Supported destination schemes:
- `rsync://user:pass@host/module/path`
- `smb://user:pass@host/share/path` (domain user example: `smb://DOMAIN%5Cuser:pass@host/share/path`)
- `davfs://user:pass@host/path` (WebDAV over HTTPS)
- `sftp://user:pass@host/path`
- Username/password components should be URL-encoded when they contain reserved characters (`@`, `:`, `/`, `\`, `%`, `#`, `?`).
- Example:
```env
BACKUP_DESTINATION=sftp://backupuser:StrongPassword@sftp.example.com/exports/samba
```
## Useful Commands
@@ -178,6 +195,7 @@ docker compose exec samba python3 /app/reconcile_shares.py
docker compose exec samba sqlite3 /state/shares.db 'SELECT * FROM shares;'
docker compose exec samba testparm -s
docker compose exec samba sh -lc 'tail -n 200 /var/log/samba/log.*'
docker compose exec samba sh -lc 'tail -n 200 /var/log/backup.log'
```
## Troubleshooting
@@ -213,9 +231,9 @@ docker compose exec samba sh -lc 'tail -n 200 /var/log/samba/log.*'
docker compose exec samba wbinfo -g
```
### Dynamic shares not appearing
### Data folders not appearing
- Confirm AD groups match `FileShare_*` or `FS_*`.
- Confirm AD groups match `FS_*`.
- Run manual reconciliation and inspect logs:
```bash
@@ -223,12 +241,6 @@ docker compose exec samba sh -lc 'tail -n 200 /var/log/samba/log.*'
docker compose exec samba tail -n 100 /var/log/reconcile.log
```
- Validate generated config:
```bash
docker compose exec samba cat /etc/samba/generated/shares.conf
```
### `acl_xattr.so` or `full_audit.so` module load error
- If logs show `Error loading module .../vfs/acl_xattr.so` (or `full_audit.so`), your running image is missing Samba VFS modules.
@@ -262,5 +274,5 @@ docker compose exec samba sh -lc 'tail -n 200 /var/log/samba/log.*'
## Notes
- User data is never automatically deleted.
- Inactive/deleted groups only stop being exposed as shares.
- Inactive/deleted FS_* groups are moved to `/data/groups/archive`.
- Data and state survive container restarts via named Docker volumes (`/data/*`, `/state`, `/var/lib/samba`).

View File

@@ -0,0 +1,322 @@
#!/usr/bin/env python3
import fcntl
import os
import subprocess
import sys
import tempfile
from dataclasses import dataclass
from typing import Dict, List, Optional, Tuple
from urllib.parse import SplitResult, unquote, urlsplit
LOCK_PATH = "/state/backup.lock"
BACKUP_SOURCES: List[Tuple[str, str]] = [
("/data/private", "data/private"),
("/data/groups", "data/groups"),
("/data/fslogix", "data/fslogix"),
("/state", "state"),
("/var/lib/samba/private", "samba/private"),
]
RCLONE_SCHEME_MAP = {
"sftp": "sftp",
"smb": "smb",
"cifs": "smb",
"davfs": "webdav",
"dav": "webdav",
"webdav": "webdav",
"http": "webdav",
"https": "webdav",
}
@dataclass
class Destination:
raw_url: str
scheme: str
parts: SplitResult
username: str
password: str
hostname: str
port: Optional[int]
path: str
def log(message: str) -> None:
print(f"[backup] {message}", flush=True)
def run_command(
command: List[str],
*,
env: Optional[Dict[str, str]] = None,
check: bool = True,
) -> subprocess.CompletedProcess:
result = subprocess.run(command, capture_output=True, text=True, env=env)
if check and result.returncode != 0:
output = result.stderr.strip() or result.stdout.strip()
raise RuntimeError(f"Command failed ({command[0]}): {output}")
return result
def parse_destination(raw_url: str) -> Destination:
parts = urlsplit(raw_url)
scheme = parts.scheme.lower()
if not scheme:
raise RuntimeError("BACKUP_DESTINATION must include a URL scheme")
hostname = parts.hostname or ""
if not hostname:
raise RuntimeError("BACKUP_DESTINATION must include a hostname")
return Destination(
raw_url=raw_url,
scheme=scheme,
parts=parts,
username=unquote(parts.username or ""),
password=unquote(parts.password or ""),
hostname=hostname,
port=parts.port,
path=unquote(parts.path or ""),
)
def redact_destination(raw_url: str) -> str:
parts = urlsplit(raw_url)
host = parts.hostname or ""
if ":" in host and not host.startswith("["):
host = f"[{host}]"
if parts.port is not None:
host = f"{host}:{parts.port}"
if parts.username:
username = unquote(parts.username)
if parts.password is not None:
auth = f"{username}:***"
else:
auth = username
netloc = f"{auth}@{host}"
else:
netloc = host
return f"{parts.scheme}://{netloc}{parts.path}"
def available_sources() -> List[Tuple[str, str]]:
sources: List[Tuple[str, str]] = []
for source_path, destination_path in BACKUP_SOURCES:
if os.path.isdir(source_path):
sources.append((source_path, destination_path))
else:
log(f"Skipping missing source: {source_path}")
return sources
def format_host(hostname: str) -> str:
if ":" in hostname and not hostname.startswith("["):
return f"[{hostname}]"
return hostname
def join_path(prefix: str, suffix: str) -> str:
left = prefix.strip("/")
right = suffix.strip("/")
if left and right:
return f"{left}/{right}"
return left or right
def obscure_secret(secret: str) -> str:
result = run_command(["rclone", "obscure", secret])
value = result.stdout.strip()
if not value:
raise RuntimeError("rclone obscure returned an empty value")
return value
def parse_smb_identity(username: str) -> Tuple[str, str]:
if not username:
return "", ""
if ";" in username:
domain, user = username.split(";", 1)
return domain, user
if "\\" in username:
domain, user = username.split("\\", 1)
return domain, user
return "", username
def sync_with_rsync(destination: Destination, sources: List[Tuple[str, str]]) -> None:
module_path = destination.path.lstrip("/")
if not module_path:
raise RuntimeError(
"rsync destinations must include a module path (example: rsync://user:pass@host/module/path)"
)
host = format_host(destination.hostname)
if destination.port is not None:
host = f"{host}:{destination.port}"
user_prefix = f"{destination.username}@" if destination.username else ""
remote_base = f"rsync://{user_prefix}{host}/{module_path.rstrip('/')}"
command_env = os.environ.copy()
if destination.password:
command_env["RSYNC_PASSWORD"] = destination.password
for source_path, destination_path in sources:
remote_path = f"{remote_base}/{destination_path.strip('/')}/"
log(f"Syncing {source_path} to rsync destination")
run_command(
["rsync", "-a", "--delete", f"{source_path}/", remote_path],
env=command_env,
)
def build_rclone_remote(destination: Destination) -> Tuple[Dict[str, str], str]:
backend = RCLONE_SCHEME_MAP.get(destination.scheme)
if backend is None:
supported = ", ".join(["rsync", *sorted(RCLONE_SCHEME_MAP.keys())])
raise RuntimeError(
f"Unsupported BACKUP_DESTINATION scheme '{destination.scheme}'. Supported schemes: {supported}"
)
options: Dict[str, str] = {"type": backend}
remote_prefix = ""
if backend == "sftp":
options["host"] = destination.hostname
if destination.port is not None:
options["port"] = str(destination.port)
if destination.username:
options["user"] = destination.username
if destination.password:
options["pass"] = obscure_secret(destination.password)
remote_prefix = destination.path.strip("/")
return options, remote_prefix
if backend == "smb":
path_segments = [segment for segment in destination.path.split("/") if segment]
if not path_segments:
raise RuntimeError(
"smb destinations must include a share name in the path (example: smb://user:pass@host/share/path)"
)
share_name = path_segments[0]
remote_prefix = "/".join(path_segments[1:])
options["host"] = destination.hostname
options["share"] = share_name
if destination.port is not None:
options["port"] = str(destination.port)
domain, user = parse_smb_identity(destination.username)
if user:
options["user"] = user
if domain:
options["domain"] = domain
if destination.password:
options["pass"] = obscure_secret(destination.password)
return options, remote_prefix
webdav_scheme = (
destination.scheme if destination.scheme in {"http", "https"} else "https"
)
host = format_host(destination.hostname)
if destination.port is not None:
host = f"{host}:{destination.port}"
webdav_path = destination.path or "/"
options["url"] = f"{webdav_scheme}://{host}{webdav_path}"
options["vendor"] = "other"
if destination.username:
options["user"] = destination.username
if destination.password:
options["pass"] = obscure_secret(destination.password)
return options, remote_prefix
def write_rclone_config(options: Dict[str, str]) -> str:
with tempfile.NamedTemporaryFile("w", encoding="utf-8", delete=False) as handle:
handle.write("[backup]\n")
for key, value in options.items():
handle.write(f"{key} = {value}\n")
config_path = handle.name
os.chmod(config_path, 0o600)
return config_path
def sync_with_rclone(destination: Destination, sources: List[Tuple[str, str]]) -> None:
options, remote_prefix = build_rclone_remote(destination)
config_path = write_rclone_config(options)
try:
for source_path, destination_path in sources:
remote_path = join_path(remote_prefix, destination_path)
log(f"Syncing {source_path} to {destination.scheme} destination")
run_command(
[
"rclone",
"sync",
f"{source_path}/",
f"backup:{remote_path}",
"--config",
config_path,
"--create-empty-src-dirs",
]
)
finally:
if os.path.exists(config_path):
os.remove(config_path)
def run_backup() -> int:
destination_url = os.getenv("BACKUP_DESTINATION", "").strip()
if not destination_url:
log("BACKUP_DESTINATION is unset, skipping backup")
return 0
sources = available_sources()
if not sources:
log("No backup sources are available, skipping backup")
return 0
destination = parse_destination(destination_url)
log(f"Starting backup to {redact_destination(destination.raw_url)}")
if destination.scheme == "rsync":
sync_with_rsync(destination, sources)
else:
sync_with_rclone(destination, sources)
log("Backup completed")
return 0
def with_lock() -> int:
lock_dir = os.path.dirname(LOCK_PATH)
if lock_dir and not os.path.isdir(lock_dir):
os.makedirs(lock_dir, exist_ok=True)
with open(LOCK_PATH, "w", encoding="utf-8") as lock_file:
try:
fcntl.flock(lock_file, fcntl.LOCK_EX | fcntl.LOCK_NB)
except BlockingIOError:
log("Another backup process is already running; skipping this cycle")
return 0
return run_backup()
def main() -> int:
try:
return with_lock()
except Exception as exc: # pylint: disable=broad-except
log(f"ERROR: {exc}")
return 1
if __name__ == "__main__":
sys.exit(main())

View File

@@ -112,15 +112,11 @@ resolve_share_groups_from_sids() {
export DOMAIN_ADMINS_GROUP
DOMAIN_ADMINS_GROUP="$(resolve_sid_to_group "$DOMAIN_ADMINS_SID")"
export PUBLIC_GROUP
PUBLIC_GROUP="$(resolve_sid_to_group "$PUBLIC_GROUP_SID")"
export FSLOGIX_GROUP
FSLOGIX_GROUP="$(resolve_sid_to_group "$FSLOGIX_GROUP_SID")"
log "Resolved DOMAIN_USERS_SID to '${DOMAIN_USERS_GROUP}'"
log "Resolved DOMAIN_ADMINS_SID to '${DOMAIN_ADMINS_GROUP}'"
log "Resolved PUBLIC_GROUP_SID to '${PUBLIC_GROUP}'"
log "Resolved FSLOGIX_GROUP_SID to '${FSLOGIX_GROUP}'"
}
@@ -159,12 +155,13 @@ write_runtime_env_file() {
printf 'export NETBIOS_NAME=%q\n' "$NETBIOS_NAME"
printf 'export DOMAIN_USERS_SID=%q\n' "$DOMAIN_USERS_SID"
printf 'export DOMAIN_ADMINS_SID=%q\n' "$DOMAIN_ADMINS_SID"
printf 'export PUBLIC_GROUP_SID=%q\n' "$PUBLIC_GROUP_SID"
printf 'export FSLOGIX_GROUP_SID=%q\n' "$FSLOGIX_GROUP_SID"
printf 'export DOMAIN_USERS_GROUP=%q\n' "$DOMAIN_USERS_GROUP"
printf 'export DOMAIN_ADMINS_GROUP=%q\n' "$DOMAIN_ADMINS_GROUP"
printf 'export PUBLIC_GROUP=%q\n' "$PUBLIC_GROUP"
printf 'export FSLOGIX_GROUP=%q\n' "$FSLOGIX_GROUP"
if [[ -n "${BACKUP_DESTINATION:-}" ]]; then
printf 'export BACKUP_DESTINATION=%q\n' "$BACKUP_DESTINATION"
fi
if [[ -n "${JOIN_USER:-}" ]]; then
printf 'export JOIN_USER=%q\n' "$JOIN_USER"
fi
@@ -217,6 +214,13 @@ SHELL=/bin/bash
PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
*/5 * * * * root source /app/runtime.env && /usr/bin/python3 /app/reconcile_shares.py >> /var/log/reconcile.log 2>&1
EOF
if [[ -n "${BACKUP_DESTINATION:-}" ]]; then
cat >> /etc/cron.d/reconcile-shares <<'EOF'
*/30 * * * * root source /app/runtime.env && /usr/bin/python3 /app/backup_to_destination.py >> /var/log/backup.log 2>&1
EOF
fi
chmod 0644 /etc/cron.d/reconcile-shares
}
@@ -227,11 +231,9 @@ require_env DOMAIN_USERS_SID
require_env DOMAIN_ADMINS_SID
export REALM WORKGROUP DOMAIN
export PUBLIC_GROUP_SID="${PUBLIC_GROUP_SID:-${DOMAIN_USERS_SID}}"
export FSLOGIX_GROUP_SID="${FSLOGIX_GROUP_SID:-${DOMAIN_USERS_SID}}"
export DOMAIN_USERS_GROUP="${DOMAIN_USERS_SID}"
export DOMAIN_ADMINS_GROUP="${DOMAIN_ADMINS_SID}"
export PUBLIC_GROUP="${PUBLIC_GROUP_SID}"
export FSLOGIX_GROUP="${FSLOGIX_GROUP_SID}"
if [[ -n "${JOIN_USER:-}" ]]; then
export JOIN_USER
@@ -240,8 +242,8 @@ if [[ -n "${JOIN_PASSWORD:-}" ]]; then
export JOIN_PASSWORD
fi
mkdir -p /data/private /data/public /data/fslogix /data/groups /state /etc/samba/generated /var/log/samba
touch /etc/samba/generated/shares.conf /var/log/reconcile.log
mkdir -p /data/private /data/fslogix /data/groups/data /data/groups/archive /state /var/log/samba
touch /var/log/reconcile.log /var/log/backup.log
append_winbind_to_nss
require_vfs_modules
@@ -263,6 +265,15 @@ write_runtime_env_file
log 'Running startup reconciliation'
python3 /app/reconcile_shares.py
if [[ -n "${BACKUP_DESTINATION:-}" ]]; then
log 'Running startup backup'
if ! python3 /app/backup_to_destination.py; then
log 'Startup backup failed; continuing service startup.'
fi
else
log 'BACKUP_DESTINATION is unset; startup backup skipped'
fi
install_cron_job
log 'Starting cron daemon'

View File

@@ -17,22 +17,18 @@ from typing import Dict, List, Optional, Tuple
DB_PATH = "/state/shares.db"
LOCK_PATH = "/state/reconcile.lock"
GROUP_ROOT = "/data/groups"
GROUP_ROOT = "/data/groups/data"
GROUP_ARCHIVE_ROOT = "/data/groups/archive"
PRIVATE_ROOT = "/data/private"
PUBLIC_ROOT = "/data/public"
FSLOGIX_ROOT = "/data/fslogix"
GENERATED_CONF = "/etc/samba/generated/shares.conf"
LDAP_FILTER = (
"(&(objectClass=group)(|(sAMAccountName=FileShare_*)(sAMAccountName=FS_*)))"
)
GROUP_PREFIXES = ("FileShare_", "FS_")
LDAP_FILTER = "(&(objectClass=group)(sAMAccountName=FS_*))"
GROUP_PREFIXES = ("FS_",)
USER_STATUS_FILTER = "(&(objectClass=user)(!(objectClass=computer))(sAMAccountName=*))"
GROUP_TITLE_ATTRS = ("displayname", "name", "cn")
REQUIRED_ENV = ["REALM", "WORKGROUP", "DOMAIN"]
ATTR_RE = re.compile(r"^([^:]+)(::?)\s*(.*)$")
SHARE_NAME_INVALID_RE = re.compile(r"[\\/:*?\"<>|;\[\],+=]")
GROUP_FOLDER_INVALID_RE = re.compile(r"[\\/:*?\"<>|;\[\],+=]")
PRIVATE_SKIP_EXACT = {
"krbtgt",
"administrator",
@@ -47,6 +43,7 @@ PRIVATE_SKIP_PREFIXES = ("msol_", "fileshare_service", "aad_")
UAC_ACCOUNTDISABLE = 0x0002
UAC_LOCKOUT = 0x0010
AD_NEVER_EXPIRES_VALUES = {0, 9223372036854775807}
MAX_GROUP_FOLDER_NAME = 120
def now_utc() -> str:
@@ -125,15 +122,6 @@ def derive_share_name(sam_account_name: str) -> Optional[str]:
return None
def derive_group_title(entry: Dict[str, Tuple[str, bool]]) -> Optional[str]:
for attr in GROUP_TITLE_ATTRS:
if attr in entry:
value = entry[attr][0].strip()
if value:
return value
return None
def parse_groups_from_ldap_output(output: str) -> List[Dict[str, str]]:
entries = parse_ldap_entries(output)
@@ -144,7 +132,7 @@ def parse_groups_from_ldap_output(output: str) -> List[Dict[str, str]]:
sam_value, _ = entry["samaccountname"]
sam = sam_value.strip()
share_name = derive_group_title(entry) or derive_share_name(sam)
share_name = derive_share_name(sam)
if not share_name:
continue
@@ -166,20 +154,76 @@ def parse_groups_from_ldap_output(output: str) -> List[Dict[str, str]]:
return list(deduped.values())
def sanitize_group_folder_name(raw_name: str) -> str:
candidate = GROUP_FOLDER_INVALID_RE.sub("_", raw_name.strip())
candidate = candidate.strip().strip(".")
candidate = re.sub(r"\s+", " ", candidate)
if not candidate:
return ""
return candidate[:MAX_GROUP_FOLDER_NAME]
def with_suffix_limited(base_name: str, suffix: str) -> str:
limit = MAX_GROUP_FOLDER_NAME - len(suffix)
if limit <= 0:
return suffix[-MAX_GROUP_FOLDER_NAME:]
trimmed = base_name[:limit].rstrip(" ._")
if not trimmed:
trimmed = "group"
return f"{trimmed}{suffix}"
def choose_group_folder_name(
preferred_name: str,
sam_account_name: str,
guid: str,
used_names: set,
existing_name: str = "",
) -> str:
base_name = sanitize_group_folder_name(preferred_name)
if not base_name:
base_name = sanitize_group_folder_name(
derive_share_name(sam_account_name) or ""
)
if not base_name:
base_name = sanitize_group_folder_name(sam_account_name)
if not base_name:
base_name = f"group_{guid[:8]}"
if existing_name:
existing_folded = existing_name.casefold()
if existing_folded not in used_names:
sanitized_existing = sanitize_group_folder_name(existing_name)
if sanitized_existing.casefold() == base_name.casefold():
used_names.add(existing_folded)
return existing_name
candidate = base_name
index = 0
while candidate.casefold() in used_names:
index += 1
suffix = f"_{guid[:8]}" if index == 1 else f"_{guid[:8]}_{index}"
candidate = with_suffix_limited(base_name, suffix)
used_names.add(candidate.casefold())
return candidate
def next_available_path(path: str) -> str:
if not os.path.exists(path):
return path
index = 1
while True:
candidate = f"{path}_{index}"
if not os.path.exists(candidate):
return candidate
index += 1
def fetch_groups_via_net_ads() -> List[Dict[str, str]]:
result = run_command(
[
"net",
"ads",
"search",
"-P",
LDAP_FILTER,
"objectGUID",
"sAMAccountName",
"displayName",
"name",
"cn",
],
["net", "ads", "search", "-P", LDAP_FILTER, "objectGUID", "sAMAccountName"],
check=False,
)
if result.returncode != 0:
@@ -226,9 +270,6 @@ def fetch_groups_via_ldap_bind() -> List[Dict[str, str]]:
LDAP_FILTER,
"objectGUID",
"sAMAccountName",
"displayName",
"name",
"cn",
]
)
return parse_groups_from_ldap_output(result.stdout)
@@ -334,143 +375,103 @@ def reconcile_db(conn: sqlite3.Connection, ad_groups: List[Dict[str, str]]) -> N
timestamp = now_utc()
seen = set()
os.makedirs(GROUP_ROOT, exist_ok=True)
os.makedirs(GROUP_ARCHIVE_ROOT, exist_ok=True)
rows = conn.execute(
"SELECT objectGUID, samAccountName, shareName, path, isActive FROM shares"
).fetchall()
existing_by_guid = {row["objectGUID"]: row for row in rows}
used_names = set()
for group in ad_groups:
guid = group["objectGUID"]
sam = group["samAccountName"]
share_name = group["shareName"]
row = existing_by_guid.get(guid)
existing_name = ""
if (
row is not None
and row["isActive"]
and row["path"].startswith(f"{GROUP_ROOT}/")
):
existing_name = os.path.basename(row["path"])
folder_name = choose_group_folder_name(
group["shareName"], sam, guid, used_names, existing_name
)
path = os.path.join(GROUP_ROOT, folder_name)
seen.add(guid)
row = conn.execute(
"SELECT objectGUID, path FROM shares WHERE objectGUID = ?", (guid,)
).fetchone()
if row is None:
path = os.path.join(GROUP_ROOT, guid)
ensure_group_path(path)
conn.execute(
"""
INSERT INTO shares (objectGUID, samAccountName, shareName, path, createdAt, lastSeenAt, isActive)
VALUES (?, ?, ?, ?, ?, ?, 1)
""",
(guid, sam, share_name, path, timestamp, timestamp),
(guid, sam, folder_name, path, timestamp, timestamp),
)
log(f"Discovered new share group {sam} ({guid})")
log(f"Discovered new data folder group {sam} ({guid})")
continue
path = row["path"]
existing_path = row["path"]
if existing_path != path:
if os.path.exists(existing_path):
final_path = next_available_path(path)
os.rename(existing_path, final_path)
path = final_path
else:
ensure_group_path(path)
ensure_group_path(path)
conn.execute(
"""
UPDATE shares
SET samAccountName = ?,
shareName = ?,
path = ?,
lastSeenAt = ?,
isActive = 1
WHERE objectGUID = ?
""",
(sam, share_name, timestamp, guid),
(sam, folder_name, path, timestamp, guid),
)
if seen:
placeholders = ",".join("?" for _ in seen)
conn.execute(
f"UPDATE shares SET isActive = 0, lastSeenAt = ? WHERE isActive = 1 AND objectGUID NOT IN ({placeholders})",
(timestamp, *seen),
active_rows = conn.execute(
"SELECT objectGUID, samAccountName, shareName, path FROM shares WHERE isActive = 1"
).fetchall()
for row in active_rows:
guid = row["objectGUID"]
if guid in seen:
continue
old_path = row["path"]
archive_name = choose_group_folder_name(
row["shareName"],
row["samAccountName"],
guid,
set(),
)
else:
archive_path = os.path.join(GROUP_ARCHIVE_ROOT, archive_name)
if os.path.exists(old_path):
final_archive_path = next_available_path(archive_path)
os.rename(old_path, final_archive_path)
archive_path = final_archive_path
conn.execute(
"UPDATE shares SET isActive = 0, lastSeenAt = ? WHERE isActive = 1",
(timestamp,),
"""
UPDATE shares
SET isActive = 0,
path = ?,
lastSeenAt = ?
WHERE objectGUID = ?
""",
(archive_path, timestamp, guid),
)
conn.commit()
def qualify_group(group_name: str) -> str:
if "\\" in group_name:
return f'@"{group_name}"'
workgroup = os.getenv("WORKGROUP", "").strip()
if workgroup:
return f'@"{workgroup}\\{group_name}"'
return f'@"{group_name}"'
def is_valid_share_name(share_name: str) -> bool:
if not share_name or share_name.casefold() in {"global", "homes", "printers"}:
return False
if SHARE_NAME_INVALID_RE.search(share_name):
return False
return True
def render_dynamic_shares(conn: sqlite3.Connection) -> None:
rows = conn.execute(
"""
SELECT objectGUID, samAccountName, shareName, path
FROM shares
WHERE isActive = 1
ORDER BY shareName COLLATE NOCASE
"""
).fetchall()
stanzas: List[str] = [
"# This file is generated by /app/reconcile_shares.py.",
"# Manual changes will be overwritten.",
"",
]
used_share_names = set()
for row in rows:
share_name = row["shareName"].strip()
if not share_name:
continue
folded_name = share_name.casefold()
if folded_name in used_share_names:
log(
f"Skipping duplicate share name '{share_name}' for objectGUID {row['objectGUID']}"
)
continue
if not is_valid_share_name(share_name):
log(
f"Skipping invalid SMB share name '{share_name}' for objectGUID {row['objectGUID']}"
)
continue
used_share_names.add(folded_name)
valid_users = qualify_group(row["samAccountName"])
stanzas.extend(
[
f"[{share_name}]",
f"path = {row['path']}",
"read only = no",
"browseable = yes",
"guest ok = no",
"vfs objects = acl_xattr full_audit",
"full_audit:prefix = %T|%u|%I|%m|%S",
"full_audit:success = all",
"full_audit:failure = all",
"full_audit:syslog = false",
f"valid users = {valid_users}",
"create mask = 0660",
"directory mask = 2770",
"inherit permissions = yes",
"access based share enum = yes",
"",
]
)
content = "\n".join(stanzas).rstrip() + "\n"
os.makedirs(os.path.dirname(GENERATED_CONF), exist_ok=True)
with tempfile.NamedTemporaryFile(
"w", encoding="utf-8", dir=os.path.dirname(GENERATED_CONF), delete=False
) as tmp_file:
tmp_file.write(content)
temp_path = tmp_file.name
os.replace(temp_path, GENERATED_CONF)
def reload_samba() -> None:
result = run_command(["smbcontrol", "all", "reload-config"], check=False)
if result.returncode != 0:
@@ -695,27 +696,6 @@ def should_skip_private_user(username: str) -> bool:
return False
def sync_public_directory() -> None:
workgroup = os.environ["WORKGROUP"]
public_group = os.getenv("PUBLIC_GROUP", "")
public_group_sid = os.getenv("PUBLIC_GROUP_SID", "")
qualified_group = public_group
os.makedirs(PUBLIC_ROOT, exist_ok=True)
gid = None
if qualified_group:
gid = resolve_group_gid_flexible(workgroup, qualified_group)
if gid is None and public_group_sid:
gid = resolve_gid_from_sid(public_group_sid)
if gid is not None:
admin_gid = resolve_gid_from_sid(os.getenv("DOMAIN_ADMINS_SID", ""))
enforce_group_tree_permissions(PUBLIC_ROOT, gid, admin_gid)
else:
group_display = qualified_group or public_group_sid or "<unset>"
log(f"Unable to resolve GID for {group_display}; public ACLs unchanged")
def sync_fslogix_directory() -> None:
workgroup = os.environ["WORKGROUP"]
fslogix_group = os.getenv("FSLOGIX_GROUP", "")
@@ -820,6 +800,11 @@ def sync_dynamic_directory_permissions(conn: sqlite3.Connection) -> None:
enforce_group_tree_permissions(path, gid, admin_gid)
os.makedirs(GROUP_ROOT, exist_ok=True)
os.chown(GROUP_ROOT, 0, 0)
run_command(["setfacl", "-b", GROUP_ROOT], check=False)
os.chmod(GROUP_ROOT, 0o555)
def with_lock() -> bool:
os.makedirs(os.path.dirname(LOCK_PATH), exist_ok=True)
@@ -834,18 +819,17 @@ def with_lock() -> bool:
try:
ensure_required_env()
os.makedirs(GROUP_ROOT, exist_ok=True)
os.makedirs(GROUP_ARCHIVE_ROOT, exist_ok=True)
conn = open_db()
try:
groups = fetch_fileshare_groups()
log(f"Discovered {len(groups)} dynamic share group(s) from AD")
log(f"Discovered {len(groups)} data folder group(s) from AD")
reconcile_db(conn, groups)
sync_dynamic_directory_permissions(conn)
render_dynamic_shares(conn)
finally:
conn.close()
sync_public_directory()
sync_fslogix_directory()
sync_private_directories()
reload_samba()

View File

@@ -14,7 +14,6 @@ services:
- "139:139"
volumes:
- private_data:/data/private
- public_data:/data/public
- fslogix_data:/data/fslogix
- group_data:/data/groups
- state_data:/state
@@ -28,7 +27,6 @@ services:
volumes:
private_data:
public_data:
fslogix_data:
group_data:
state_data:

View File

@@ -37,8 +37,6 @@
logging = file
log level = 1 auth:5 passdb:5 winbind:3
include = /etc/samba/generated/shares.conf
[Privat]
path = /data/private
read only = no
@@ -55,8 +53,8 @@
ea support = yes
nt acl support = no
[Geteilt]
path = /data/public
[Data]
path = /data/groups/data
read only = no
browseable = yes
guest ok = no
@@ -65,11 +63,11 @@
full_audit:success = all
full_audit:failure = all
full_audit:syslog = false
valid users = @"${PUBLIC_GROUP}"
force group = "${PUBLIC_GROUP}"
valid users = @"${DOMAIN_USERS_GROUP}"
create mask = 0660
directory mask = 2770
inherit permissions = yes
hide unreadable = yes
access based share enum = yes
[FSLogix]

19
setup
View File

@@ -83,13 +83,12 @@ write_env_file() {
local admin_password=""
local domain_users_sid=""
local domain_admins_sid=""
local public_group_sid=""
local fslogix_group_sid=""
local backup_destination=""
local samba_hostname="adsambafsrv"
local netbios_name="ADSAMBAFSRV"
local service_password=""
local service_account_sam=""
local public_group_prompt=""
local fslogix_group_prompt=""
local samba_hostname_input=""
local netbios_name_input=""
@@ -103,12 +102,6 @@ write_env_file() {
prompt_value domain_users_sid "DOMAIN_USERS_SID (e.g. ...-513)"
prompt_value domain_admins_sid "DOMAIN_ADMINS_SID (e.g. ...-512)"
public_group_prompt="PUBLIC_GROUP_SID (press Enter to reuse DOMAIN_USERS_SID)"
read -r -p "${public_group_prompt}: " public_group_sid
if [[ -z "$public_group_sid" ]]; then
public_group_sid="$domain_users_sid"
fi
fslogix_group_prompt="FSLOGIX_GROUP_SID (press Enter to reuse DOMAIN_USERS_SID)"
read -r -p "${fslogix_group_prompt}: " fslogix_group_sid
if [[ -z "$fslogix_group_sid" ]]; then
@@ -130,6 +123,8 @@ write_env_file() {
fi
netbios_name="$sanitized_netbios_name"
read -r -p "BACKUP_DESTINATION (optional URL, press Enter to disable): " backup_destination
service_account_sam="$(sanitize_sam_account_name "$SERVICE_ACCOUNT_NAME")"
if [[ "$service_account_sam" != "$SERVICE_ACCOUNT_NAME" ]]; then
printf "Using sAMAccountName '%s' (AD limit is 20 chars; requested '%s').\n" "$service_account_sam" "$SERVICE_ACCOUNT_NAME"
@@ -165,8 +160,8 @@ SERVICE_ACCOUNT_SAM=${service_account_sam}
SERVICE_ACCOUNT_PASSWORD=${service_password}
DOMAIN_USERS_SID=${domain_users_sid}
DOMAIN_ADMINS_SID=${domain_admins_sid}
PUBLIC_GROUP_SID=${public_group_sid}
FSLOGIX_GROUP_SID=${fslogix_group_sid}
BACKUP_DESTINATION=${backup_destination}
SAMBA_HOSTNAME=${samba_hostname}
NETBIOS_NAME=${netbios_name}
EOF
@@ -217,8 +212,8 @@ JOIN_USER=${service_account_sam}
JOIN_PASSWORD=${service_password}
DOMAIN_USERS_SID=${domain_users_sid}
DOMAIN_ADMINS_SID=${domain_admins_sid}
PUBLIC_GROUP_SID=${public_group_sid}
FSLOGIX_GROUP_SID=${fslogix_group_sid}
BACKUP_DESTINATION=${backup_destination}
SAMBA_HOSTNAME=${samba_hostname}
NETBIOS_NAME=${netbios_name}
# Optional overrides:
@@ -226,6 +221,10 @@ NETBIOS_NAME=${netbios_name}
# LDAP_BASE_DN=DC=example,DC=com
# PRIVATE_SKIP_USERS=svc_backup,svc_sql
# PRIVATE_SKIP_PREFIXES=svc_,sql_
# BACKUP_DESTINATION=rsync://user:pass@backup.example.com/samba-backups
# BACKUP_DESTINATION=smb://DOMAIN%5Cuser:pass@backup.example.com/Backups/samba
# BACKUP_DESTINATION=davfs://user:pass@webdav.example.com/remote.php/dav/files/backup
# BACKUP_DESTINATION=sftp://user:pass@sftp.example.com/exports/samba
EOF
chmod 600 "$ENV_FILE"