for last 7 days of messages remove large messages first, then by age

Both dovecot-quota-threshold triggers and the daily expiry routine
will now expunge oldest messages from mailboxes automatically
when the mailbox reaches 75% of max_mailbox_size.
Delta Chat users should not see any warnings (at 80/95 percent) or bounce messages,
and existing over-quota mailboxes should start receiving mails again.

chatmaild/pyproject.toml

@@ -21,7 +21,8 @@ where = ['src']
 [project.scripts]
 doveauth = "chatmaild.doveauth:main"
 chatmail-metadata = "chatmaild.metadata:main"
-chatmail-expire = "chatmaild.expire:main"
+chatmail-expire = "chatmaild.expire:daily_expire_main"
+chatmail-quota-expire = "chatmaild.expire:quota_expire_main"
 chatmail-fsreport = "chatmaild.fsreport:main"
 lastlogin = "chatmaild.lastlogin:main"
 turnserver = "chatmaild.turnserver:main"

chatmaild/src/chatmaild/config.py

@@ -95,6 +95,11 @@ class Config:
         # old unused option (except for first migration from sqlite to maildir store)
         self.passdb_path = Path(params.get("passdb_path", "/home/vmail/passdb.sqlite"))
 
+    @property
+    def max_mailbox_size_mb(self):
+        """Return max_mailbox_size as an integer in megabytes."""
+        return parse_size_mb(self.max_mailbox_size)
+
     def _getbytefile(self):
         return open(self._inipath, "rb")
 
@@ -108,6 +113,16 @@ class Config:
         return User(maildir, addr, password_path, uid="vmail", gid="vmail")
 
 
+def parse_size_mb(limit):
+    """Parse a size string like ``500M`` or ``2G`` and return megabytes."""
+    value = limit.strip().upper().removesuffix("B")
+    if value.endswith("G"):
+        return int(value[:-1]) * 1024
+    if value.endswith("M"):
+        return int(value[:-1])
+    return int(value)
+
+
 def write_initial_config(inipath, mail_domain, overrides):
     """Write out default config file, using the specified config value overrides."""
     content = get_default_config_content(mail_domain, **overrides)

chatmaild/src/chatmaild/expire.py

@@ -4,17 +4,26 @@ Expire old messages and addresses.
 """
 
 import os
+import re
 import shutil
 import sys
 import time
 from argparse import ArgumentParser
 from collections import namedtuple
 from datetime import datetime
+from pathlib import Path
 from stat import S_ISREG
 
 from chatmaild.config import read_config
 
 FileEntry = namedtuple("FileEntry", ("path", "mtime", "size"))
+QuotaFileEntry = namedtuple("QuotaFileEntry", ("mtime", "quota_size", "path"))
+
+# Quota cleanup factor of max_mailbox_size. The mailbox is reset to this size.
+QUOTA_CLEANUP_FACTOR = 0.7
+
+# e.g. "cur/1775324677.M448978P3029757.exam,S=3235,W=3305:2,S"
+_dovecot_fn_rex = re.compile(r".+/(\d+)\..+,S=(\d+)")
 
 
 def iter_mailboxes(basedir, maxnum):
@@ -74,6 +83,54 @@ class MailboxStat:
         self.extrafiles.sort(key=lambda x: -x.size)
 
 
+def parse_dovecot_filename(relpath):
+    m = _dovecot_fn_rex.match(relpath)
+    if not m:
+        return None
+    return QuotaFileEntry(int(m.group(1)), int(m.group(2)), relpath)
+
+
+def scan_mailbox_messages(mbox):
+    messages = []
+    for sub in ("cur", "new"):
+        for name in os_listdir_if_exists(mbox / sub):
+            if entry := parse_dovecot_filename(f"{sub}/{name}"):
+                messages.append(entry)
+    return messages
+
+
+# Within this window, large messages are deleted before small ones.
+DELETE_LARGE_FIRST_DAYS = 7
+
+
+def expire_to_target(mbox, target_bytes):
+    cutoff = time.time() - DELETE_LARGE_FIRST_DAYS * 86400
+    messages = scan_mailbox_messages(mbox)
+
+    def sort_key(msg):
+        # prio 0: Older than cutoff -> remove oldest first
+        if msg.mtime < cutoff:
+            return (0, msg.mtime)
+
+        # prio 1: more recent than cutoff, large -> remove largest first
+        if msg.quota_size > 200000:
+            return (1, -msg.quota_size, msg.mtime)
+
+        # prio 2: more recent than cutoff, small -> remove oldest first
+        return (2, msg.mtime)
+
+    total_size = sum(m.quota_size for m in messages)
+    removed = 0
+    for entry in sorted(messages, key=sort_key):
+        if total_size <= target_bytes:
+            break
+        (mbox / entry.path).unlink(missing_ok=True)
+        total_size -= entry.quota_size
+        removed += 1
+
+    return removed
+
+
 def print_info(msg):
     print(msg, file=sys.stderr)
 
@@ -143,6 +200,19 @@ class Expiry:
             else:
                 continue
             changed = True
+
+        target_bytes = (
+            self.config.max_mailbox_size_mb * 1024 * 1024 * QUOTA_CLEANUP_FACTOR
+        )
+        removed = expire_to_target(Path(mbox.basedir), target_bytes)
+        if removed:
+            changed = True
+            self.del_files += removed
+            if self.verbose:
+                print_info(
+                    f"quota-expire: removed {removed} message(s) from {mboxname}"
+                )
+
         if changed:
             self.remove_file(f"{mbox.basedir}/maildirsize")
 
@@ -154,9 +224,9 @@ class Expiry:
         )
 
 
-def main(args=None):
+def daily_expire_main(args=None):
     """Expire mailboxes and messages according to chatmail config"""
-    parser = ArgumentParser(description=main.__doc__)
+    parser = ArgumentParser(description=daily_expire_main.__doc__)
     ini = "/usr/local/lib/chatmaild/chatmail.ini"
     parser.add_argument(
         "chatmail_ini",
@@ -202,5 +272,33 @@ def main(args=None):
     print(exp.get_summary())
 
 
-if __name__ == "__main__":
-    main(sys.argv[1:])
+def quota_expire_main(args=None):
+    """Remove mailbox messages to stay within a megabyte target.
+
+    This entry point is called by dovecot when a quota threshold is passed.
+    """
+
+    parser = ArgumentParser(description=quota_expire_main.__doc__)
+    parser.add_argument(
+        "target_mb",
+        type=int,
+        help="target mailbox size in megabytes",
+    )
+    parser.add_argument(
+        "mailbox_path",
+        type=Path,
+        help="path to a user mailbox",
+    )
+    args = parser.parse_args(args)
+
+    target_bytes = args.target_mb * 1024 * 1024
+
+    removed_count = expire_to_target(args.mailbox_path, target_bytes)
+    if removed_count:
+        (args.mailbox_path / "maildirsize").unlink(missing_ok=True)
+        print(
+            f"quota-expire: removed {removed_count} message(s)"
+            f" from {args.mailbox_path.name}",
+            file=sys.stderr,
+        )
+    return 0

chatmaild/src/chatmaild/ini/chatmail.ini.f

@@ -18,6 +18,7 @@ max_user_send_per_minute = 60
 max_user_send_burst_size = 10
 
 # maximum mailbox size of a chatmail address
+# Oldest messages will be removed automatically, so mailboxes never run full.
 max_mailbox_size = 500M
 
 # maximum message size for an e-mail in bytes

chatmaild/src/chatmaild/tests/test_config.py

@@ -1,6 +1,6 @@
 import pytest
 
-from chatmaild.config import read_config
+from chatmaild.config import parse_size_mb, read_config
 
 
 def test_read_config_basic(example_config):
@@ -121,3 +121,17 @@ def test_config_tls_external_bad_format(make_config):
                 "tls_external_cert_and_key": "/only/one/path.pem",
             },
         )
+
+
+def test_parse_size_mb():
+    assert parse_size_mb("500M") == 500
+    assert parse_size_mb("2G") == 2048
+    assert parse_size_mb("  1g  ") == 1024
+    assert parse_size_mb("100MB") == 100
+    assert parse_size_mb("256") == 256
+
+
+def test_max_mailbox_size_mb(make_config):
+    config = make_config("chat.example.org")
+    assert config.max_mailbox_size == "500M"
+    assert config.max_mailbox_size_mb == 500

chatmaild/src/chatmaild/tests/test_delete_inactive_users.py

@@ -1,7 +1,7 @@
 import time
 
 from chatmaild.doveauth import AuthDictProxy
-from chatmaild.expire import main as main_expire
+from chatmaild.expire import daily_expire_main as main_expire
 
 
 def test_login_timestamps(example_config):

chatmaild/src/chatmaild/tests/test_expire.py

@@ -1,5 +1,8 @@
+import itertools
 import os
 import random
+import shutil
+import time
 from datetime import datetime
 from fnmatch import fnmatch
 from pathlib import Path
@@ -9,13 +12,19 @@ import pytest
 from chatmaild.expire import (
     FileEntry,
     MailboxStat,
+    expire_to_target,
     get_file_entry,
     iter_mailboxes,
     os_listdir_if_exists,
+    parse_dovecot_filename,
+    quota_expire_main,
+    scan_mailbox_messages,
 )
-from chatmaild.expire import main as expiry_main
+from chatmaild.expire import daily_expire_main as expiry_main
 from chatmaild.fsreport import main as report_main
 
+MB = 1024 * 1024
+
 
 def fill_mbox(folderdir):
     password = folderdir.joinpath("password")
@@ -196,3 +205,74 @@ def test_os_listdir_if_exists(tmp_path):
     tmp_path.joinpath("x").write_text("hello")
     assert len(os_listdir_if_exists(str(tmp_path))) == 1
     assert len(os_listdir_if_exists(str(tmp_path.joinpath("123123")))) == 0
+
+
+# --- quota expire tests ---
+
+_msg_counter = itertools.count(1)
+
+
+def _create_message(basedir, sub, size, days_old=0, disk_size=None):
+    seq = next(_msg_counter)
+    mtime = int(time.time() - days_old * 86400)
+    name = f"{mtime}.M1P1Q{seq}.hostname,S={size},W={size}:2,S"
+    path = basedir / sub / name
+    path.parent.mkdir(parents=True, exist_ok=True)
+    path.write_bytes(b"x" * (disk_size if disk_size is not None else size))
+    os.utime(path, (mtime, mtime))
+    return path
+
+
+def test_parse_dovecot_filename():
+    e = parse_dovecot_filename("cur/1775324677.M448978P3029757.exam,S=3235,W=3305:2,S")
+    assert e.path == "cur/1775324677.M448978P3029757.exam,S=3235,W=3305:2,S"
+    assert e.mtime == 1775324677
+    assert e.quota_size == 3235
+    assert parse_dovecot_filename("cur/msg_without_structure") is None
+
+
+def test_expire_to_target(tmp_path):
+    _create_message(tmp_path, "cur", MB, days_old=10, disk_size=100)
+    _create_message(tmp_path, "new", MB, days_old=5)
+    assert len(scan_mailbox_messages(tmp_path)) == 2
+    # removes oldest first, uses S= size not disk size
+    removed = expire_to_target(tmp_path, MB)
+    assert removed == 1
+    assert len(scan_mailbox_messages(tmp_path)) == 1
+
+
+def test_expire_to_target_prioritization(tmp_path):
+    def create_messages():
+        for sub in ("cur", "new"):
+            if (tmp_path / sub).exists():
+                shutil.rmtree(tmp_path / sub)
+        # prio 0: older than 7 days
+        _create_message(tmp_path, "cur", 5 * MB, days_old=10)
+        # prio 1: last 7 days, large (>200KB)
+        _create_message(tmp_path, "cur", 5 * MB, days_old=1)
+        # prio 2: last 7 days, small
+        _create_message(tmp_path, "cur", 1000, days_old=2)
+
+    # Shrink to 6MB: only the old message (prio 0) is removed.
+    create_messages()
+    assert expire_to_target(tmp_path, 6 * MB) == 1
+    msgs = scan_mailbox_messages(tmp_path)
+    assert len(msgs) == 2
+    assert all(m.mtime > time.time() - 7 * 86400 for m in msgs)
+
+    # Shrink to 1KB: old and recent-large removed, small survives.
+    create_messages()
+    assert expire_to_target(tmp_path, 1024) == 2
+    msgs = scan_mailbox_messages(tmp_path)
+    assert len(msgs) == 1
+    assert msgs[0].quota_size == 1000
+
+
+def test_quota_expire_main(tmp_path, capsys):
+    mbox = tmp_path / "user@example.org"
+    _create_message(mbox, "cur", 2 * MB, days_old=5)
+    (mbox / "maildirsize").write_text("x")
+    quota_expire_main([str(1), str(mbox)])
+    _, err = capsys.readouterr()
+    assert "quota-expire: removed 1 message(s) from user@example.org" in err
+    assert not (mbox / "maildirsize").exists()

cmdeploy/src/cmdeploy/dovecot/dovecot.conf.j2

@@ -149,12 +149,26 @@ plugin {
 }
 
 plugin {
-  # for now we define static quota-rules for all users 
   quota = maildir:User quota
-  quota_rule = *:storage={{ config.max_mailbox_size }}
   quota_max_mail_size={{ config.max_message_size }}
   quota_grace = 0
-  # quota_over_flag_value = TRUE
+
+  quota_rule = *:storage={{ config.max_mailbox_size_mb }}M
+
+  # Trigger at 75%% of quota, expire oldest messages down to 70%%.
+  # The percentages are chosen to prevent current Delta Chat users
+  # from seeing "quota warnings" which trigger at 80% and 95%.
+
+  quota_warning = storage=75%% quota-warning {{ config.max_mailbox_size_mb * 70 // 100 }} {{ config.mailboxes_dir }}/%u
+}
+
+service quota-warning {
+  executable = script /usr/local/lib/chatmaild/venv/bin/chatmail-quota-expire
+  user = vmail
+  unix_listener quota-warning {
+     user = vmail
+     mode = 0600
+  }
 }
 
 # push_notification configuration

cmdeploy/src/cmdeploy/mtail/delivered_mail.mtail

@@ -78,3 +78,11 @@ counter rejected_unencrypted_mail_count
 /Rejected unencrypted mail/ {
   rejected_unencrypted_mail_count++
 }
+
+counter quota_expire_runs
+counter quota_expire_removed_files
+
+/quota-expire: removed (?P<count>\d+) message\(s\)/ {
+  quota_expire_runs++
+  quota_expire_removed_files += $count
+}

doc/source/overview.rst

@@ -102,8 +102,12 @@ short overview of ``chatmaild`` services:
    Apple/Google/Huawei.
 
 -  `chatmail-expire <https://github.com/chatmail/relay/blob/main/chatmaild/src/chatmaild/expire.py>`_
-   deletes users if they have not logged in for a longer while.
-   The timeframe can be configured in ``chatmail.ini``.
+   deletes old messages, large messages, and entire mailboxes
+   of users who have not logged in for longer than
+   ``delete_inactive_users_after`` days.
+
+-  ``chatmail-quota-expire`` is called by Dovecot's ``quota_warning`` mechanism
+   and will automatically remove oldest messages to keep mailboxes well under ``max_mailbox_size``.
 
 -  `lastlogin <https://github.com/chatmail/relay/blob/main/chatmaild/src/chatmaild/lastlogin.py>`_
    is contacted by Dovecot when a user logs in and stores the date of
@@ -156,6 +160,8 @@ Chatmail relay dependency diagram
         /home/vmail/.../user"];
         dovecot --- |lastlogin.socket|lastlogin;
         dovecot --- chatmail-metadata;
+        dovecot --- |quota-warning|chatmail-quota-expire;
+        chatmail-quota-expire --- maildir;
         lastlogin --- maildir;
         doveauth --- maildir;
         chatmail-expire-daily --- maildir;