Mercurial > dovecot > original-hg > dovecot-1.2
view src/lib-storage/index/maildir/maildir-sync.c @ 5926:095b3adc537b HEAD
Handle mailbox deletions while syncing silently.
author | Timo Sirainen <tss@iki.fi> |
---|---|
date | Mon, 09 Jul 2007 00:52:04 +0300 |
parents | 00c5e3cbeaf0 |
children | b9865213da42 |
line wrap: on
line source
/* Copyright (C) 2004 Timo Sirainen */ /* Here's a description of how we handle Maildir synchronization and it's problems: We want to be as efficient as we can. The most efficient way to check if changes have occurred is to stat() the new/ and cur/ directories and uidlist file - if their mtimes haven't changed, there's no changes and we don't need to do anything. Problem 1: Multiple changes can happen within a single second - nothing guarantees that once we synced it, someone else didn't just then make a modification. Such modifications wouldn't get noticed until a new modification occurred later. Problem 2: Syncing cur/ directory is much more costly than syncing new/. Moving mails from new/ to cur/ will always change mtime of cur/ causing us to sync it as well. Problem 3: We may not be able to move mail from new/ to cur/ because we're out of quota, or simply because we're accessing a read-only mailbox. MAILDIR_SYNC_SECS ----------------- Several checks below use MAILDIR_SYNC_SECS, which should be maximum clock drift between all computers accessing the maildir (eg. via NFS), rounded up to next second. Our default is 1 second, since everyone should be using NTP. Note that setting it to 0 works only if there's only one computer accessing the maildir. It's practically impossible to make two clocks _exactly_ synchronized. It might be possible to only use file server's clock by looking at the atime field, but I don't know how well that would actually work. cur directory ------------- We have dirty_cur_time variable which is set to cur/ directory's mtime when it's >= time() - MAILDIR_SYNC_SECS and we _think_ we have synchronized the directory. When dirty_cur_time is non-zero, we don't synchronize the cur/ directory until a) cur/'s mtime changes b) opening a mail fails with ENOENT c) time() > dirty_cur_time + MAILDIR_SYNC_SECS This allows us to modify the maildir multiple times without having to sync it at every change. The sync will eventually be done to make sure we didn't miss any external changes. The dirty_cur_time is set when: - we change message flags - we expunge messages - we move mail from new/ to cur/ - we sync cur/ directory and it's mtime is >= time() - MAILDIR_SYNC_SECS It's unset when we do the final syncing, ie. when mtime is older than time() - MAILDIR_SYNC_SECS. new directory ------------- If new/'s mtime is >= time() - MAILDIR_SYNC_SECS, always synchronize it. dirty_cur_time-like feature might save us a few syncs, but that might break a client which saves a mail in one connection and tries to fetch it in another one. new/ directory is almost always empty, so syncing it should be very fast anyway. Actually this can still happen if we sync only new/ dir while another client is also moving mails from it to cur/ - it takes us a while to see them. That's pretty unlikely to happen however, and only way to fix it would be to always synchronize cur/ after new/. Normally we move all mails from new/ to cur/ whenever we sync it. If it's not possible for some reason, we mark the mail with "probably exists in new/ directory" flag. If rename() still fails because of ENOSPC or EDQUOT, we still save the flag changes in index with dirty-flag on. When moving the mail to cur/ directory, or when we notice it's already moved there, we apply the flag changes to the filename, rename it and remove the dirty flag. If there's dirty flags, this should be tried every time after expunge or when closing the mailbox. uidlist ------- This file contains UID <-> filename mappings. It's updated only when new mail arrives, so it may contain filenames that have already been deleted. Updating is done by getting uidlist.lock file, writing the whole uidlist into it and rename()ing it over the old uidlist. This means there's no need to lock the file for reading. Whenever uidlist is rewritten, it's mtime must be larger than the old one's. Use utime() before rename() if needed. Note that inode checking wouldn't have been sufficient as inode numbers can be reused. This file is usually read the first time you need to know filename for given UID. After that it's not re-read unless new mails come that we don't know about. broken clients -------------- Originally the middle identifier in Maildir filename was specified only as <process id>_<delivery counter>. That however created a problem with randomized PIDs which made it possible that the same PID was reused within one second. So if within one second a mail was delivered, MUA moved it to cur/ and another mail was delivered by a new process using same PID as the first one, we likely ended up overwriting the first mail when the second mail was moved over it. Nowadays everyone should be giving a bit more specific identifier, for example include microseconds in it which Dovecot does. There's a simple way to prevent this from happening in some cases: Don't move the mail from new/ to cur/ if it's mtime is >= time() - MAILDIR_SYNC_SECS. The second delivery's link() call then fails because the file is already in new/, and it will then use a different filename. There's a few problems with this however: - it requires extra stat() call which is unneeded extra I/O - another MUA might still move the mail to cur/ - if first file's flags are modified by either Dovecot or another MUA, it's moved to cur/ (you _could_ just do the dirty-flagging but that'd be ugly) Because this is useful only for very few people and it requires extra I/O, I decided not to implement this. It should be however quite easy to do since we need to be able to deal with files in new/ in any case. It's also possible to never accidentally overwrite a mail by using link() + unlink() rather than rename(). This however isn't very good idea as it introduces potential race conditions when multiple clients are accessing the mailbox: Trying to move the same mail from new/ to cur/ at the same time: a) Client 1 uses slightly different filename than client 2, for example one sets read-flag on but the other doesn't. You have the same mail duplicated now. b) Client 3 sees the mail between Client 1's and 2's link() calls and changes it's flag. You have the same mail duplicated now. And it gets worse when they're unlink()ing in cur/ directory: c) Client 1 changes mails's flag and client 2 changes it back between 1's link() and unlink(). The mail is now expunged. d) If you try to deal with the duplicates by unlink()ing another one of them, you might end up unlinking both of them. So, what should we do then if we notice a duplicate? First of all, it might not be a duplicate at all, readdir() might have just returned it twice because it was just renamed. What we should do is create a completely new base name for it and rename() it to that. If the call fails with ENOENT, it only means that it wasn't a duplicate after all. */ #include "lib.h" #include "ioloop.h" #include "array.h" #include "buffer.h" #include "hash.h" #include "str.h" #include "nfs-workarounds.h" #include "maildir-storage.h" #include "maildir-uidlist.h" #include "maildir-filename.h" #include "maildir-sync.h" #include <stdio.h> #include <stddef.h> #include <stdlib.h> #include <unistd.h> #include <dirent.h> #include <sys/stat.h> #define MAILDIR_FILENAME_FLAG_FOUND 128 /* When rename()ing many files from new/ to cur/, it's possible that next readdir() skips some files. we don't of course wish to lose them, so we go and rescan the new/ directory again from beginning until no files are left. This value is just an optimization to avoid checking the directory twice unneededly. usually only NFS is the problem case. 1 is the safest bet here, but I guess 5 will do just fine too. */ #define MAILDIR_RENAME_RESCAN_COUNT 5 /* This is mostly to avoid infinite looping when rename() destination already exists as the hard link of the file itself. */ #define MAILDIR_SCAN_DIR_MAX_COUNT 5 #define DUPE_LINKS_DELETE_SECS 30 struct maildir_sync_context { struct maildir_mailbox *mbox; const char *new_dir, *cur_dir; bool partial; time_t last_touch, last_notify; struct maildir_uidlist_sync_ctx *uidlist_sync_ctx; struct maildir_index_sync_context *index_sync_ctx; }; void maildir_sync_notify(struct maildir_sync_context *ctx) { time_t now; if (ctx == NULL) { /* we got here from maildir-save.c. it has no maildir_sync_context, */ return; } now = time(NULL); if (now - ctx->last_touch > MAILDIR_LOCK_TOUCH_SECS) { (void)maildir_uidlist_lock_touch(ctx->mbox->uidlist); ctx->last_touch = now; } if (now - ctx->last_notify > MAIL_STORAGE_STAYALIVE_SECS) { struct mailbox *box = &ctx->mbox->ibox.box; if (box->storage->callbacks->notify_ok != NULL) { box->storage->callbacks-> notify_ok(box, "Hang in there..", box->storage->callback_context); } ctx->last_notify = now; } } static struct maildir_sync_context * maildir_sync_context_new(struct maildir_mailbox *mbox) { struct maildir_sync_context *ctx; ctx = t_new(struct maildir_sync_context, 1); ctx->mbox = mbox; ctx->new_dir = t_strconcat(mbox->path, "/new", NULL); ctx->cur_dir = t_strconcat(mbox->path, "/cur", NULL); ctx->last_touch = ioloop_time; ctx->last_notify = ioloop_time; return ctx; } static void maildir_sync_deinit(struct maildir_sync_context *ctx) { if (ctx->uidlist_sync_ctx != NULL) (void)maildir_uidlist_sync_deinit(&ctx->uidlist_sync_ctx); if (ctx->index_sync_ctx != NULL) { (void)maildir_sync_index_finish(&ctx->index_sync_ctx, TRUE, FALSE); } } static int maildir_fix_duplicate(struct maildir_sync_context *ctx, const char *dir, const char *fname2) { const char *fname1, *path1, *path2; const char *new_fname, *new_path; struct stat st1, st2; int ret = 0; fname1 = maildir_uidlist_sync_get_full_filename(ctx->uidlist_sync_ctx, fname2); i_assert(fname1 != NULL); t_push(); path1 = t_strconcat(dir, "/", fname1, NULL); path2 = t_strconcat(dir, "/", fname2, NULL); if (stat(path1, &st1) < 0 || stat(path2, &st2) < 0) { /* most likely the files just don't exist anymore. don't really care about other errors much. */ t_pop(); return 0; } if (st1.st_ino == st2.st_ino && CMP_DEV_T(st1.st_dev, st2.st_dev)) { /* Files are the same. this means either a race condition between stat() calls, or that the files were link()ed. */ if (st1.st_nlink > 1 && st2.st_nlink == st1.st_nlink && st1.st_ctime == st2.st_ctime && st1.st_ctime < ioloop_time - DUPE_LINKS_DELETE_SECS) { /* The file has hard links and it hasn't had any changes (such as renames) for a while, so this isn't a race condition. rename()ing one file on top of the other would fix this safely, except POSIX decided that rename() doesn't work that way. So we'll have unlink() one and hope that another process didn't just decide to unlink() the other (uidlist lock prevents this from happening) */ if (unlink(path2) == 0) i_warning("Unlinked a duplicate: %s", path2); else { mail_storage_set_critical( &ctx->mbox->storage->storage, "unlink(%s) failed: %m", path2); } } t_pop(); return 0; } new_fname = maildir_filename_generate(); new_path = t_strconcat(ctx->mbox->path, "/new/", new_fname, NULL); if (rename(path2, new_path) == 0) i_warning("Fixed a duplicate: %s -> %s", path2, new_fname); else if (errno != ENOENT) { mail_storage_set_critical(&ctx->mbox->storage->storage, "Couldn't fix a duplicate: rename(%s, %s) failed: %m", path2, new_path); ret = -1; } t_pop(); return ret; } static int maildir_stat(struct maildir_mailbox *mbox, const char *path, struct stat *st_r) { if (nfs_safe_stat(path, st_r) == 0) return 0; if (errno == ENOENT) { /* if mailbox gets deleted under us, don't log an error */ mbox->ibox.mailbox_deleted = TRUE; return -1; } mail_storage_set_critical(mbox->ibox.box.storage, "stat(%s) failed: %m", path); return -1; } static int maildir_scan_dir(struct maildir_sync_context *ctx, bool new_dir) { struct mail_storage *storage = &ctx->mbox->storage->storage; const char *dir; DIR *dirp; string_t *src, *dest; struct dirent *dp; enum maildir_uidlist_rec_flag flags; unsigned int i = 0, move_count = 0; int ret = 1; bool move_new, check_touch; dir = new_dir ? ctx->new_dir : ctx->cur_dir; dirp = opendir(dir); if (dirp == NULL) { if (errno == ENOENT) { ctx->mbox->ibox.mailbox_deleted = TRUE; return -1; } mail_storage_set_critical(storage, "opendir(%s) failed: %m", dir); return -1; } t_push(); src = t_str_new(1024); dest = t_str_new(1024); move_new = new_dir && !mailbox_is_readonly(&ctx->mbox->ibox.box) && !ctx->mbox->ibox.keep_recent; errno = 0; for (; (dp = readdir(dirp)) != NULL; errno = 0) { if (dp->d_name[0] == '.') continue; ret = maildir_uidlist_sync_next_pre(ctx->uidlist_sync_ctx, dp->d_name); if (ret == 0) { /* new file and we couldn't lock uidlist, check this later in next sync. */ if (new_dir) ctx->mbox->last_new_mtime = 0; else ctx->mbox->dirty_cur_time = ioloop_time; continue; } if (ret < 0) break; check_touch = FALSE; flags = 0; if (move_new) { str_truncate(src, 0); str_truncate(dest, 0); str_printfa(src, "%s/%s", ctx->new_dir, dp->d_name); str_printfa(dest, "%s/%s", ctx->cur_dir, dp->d_name); if (strchr(dp->d_name, MAILDIR_INFO_SEP) == NULL) { str_append(dest, MAILDIR_FLAGS_FULL_SEP); } if (rename(str_c(src), str_c(dest)) == 0) { /* we moved it - it's \Recent for us */ move_count++; ctx->mbox->dirty_cur_time = ioloop_time; flags |= MAILDIR_UIDLIST_REC_FLAG_MOVED | MAILDIR_UIDLIST_REC_FLAG_RECENT; } else if (ENOTFOUND(errno)) { /* someone else moved it already */ move_count++; flags |= MAILDIR_UIDLIST_REC_FLAG_MOVED; } else if (ENOSPACE(errno) || errno == EACCES) { /* not enough disk space / read-only maildir, leave here */ flags |= MAILDIR_UIDLIST_REC_FLAG_NEW_DIR | MAILDIR_UIDLIST_REC_FLAG_RECENT; move_new = FALSE; } else { flags |= MAILDIR_UIDLIST_REC_FLAG_NEW_DIR | MAILDIR_UIDLIST_REC_FLAG_RECENT; mail_storage_set_critical(storage, "rename(%s, %s) failed: %m", str_c(src), str_c(dest)); } if ((move_count % MAILDIR_SLOW_MOVE_COUNT) == 0) maildir_sync_notify(ctx); } else if (new_dir) { flags |= MAILDIR_UIDLIST_REC_FLAG_NEW_DIR | MAILDIR_UIDLIST_REC_FLAG_RECENT; } i++; if ((i % MAILDIR_SLOW_CHECK_COUNT) == 0) maildir_sync_notify(ctx); ret = maildir_uidlist_sync_next(ctx->uidlist_sync_ctx, dp->d_name, flags); if (ret <= 0) { if (ret < 0) break; /* possibly duplicate - try fixing it */ if (maildir_fix_duplicate(ctx, dir, dp->d_name) < 0) { ret = -1; break; } } } if (errno != 0) { mail_storage_set_critical(storage, "readdir(%s) failed: %m", dir); ret = -1; } if (closedir(dirp) < 0) { mail_storage_set_critical(storage, "closedir(%s) failed: %m", dir); ret = -1; } t_pop(); return ret < 0 ? -1 : (move_count <= MAILDIR_RENAME_RESCAN_COUNT ? 0 : 1); } static void maildir_sync_update_from_header(struct maildir_mailbox *mbox, struct mail_index_header *hdr_r) { struct mail_index_view *view; const struct mail_index_header *hdr; /* open a new view so we get the latest header */ view = mail_index_view_open(mbox->ibox.index); hdr = mail_index_get_header(view); /* FIXME: ugly, replace with extension header */ mbox->last_new_mtime = hdr->sync_size & 0xffffffff; mbox->last_dirty_flags = (hdr->sync_size >> 32) & (MAILDIR_DIRTY_NEW | MAILDIR_DIRTY_CUR); mbox->last_cur_mtime = hdr->sync_stamp; if ((mbox->last_dirty_flags & MAILDIR_DIRTY_CUR) != 0 && mbox->dirty_cur_time < mbox->last_cur_mtime) mbox->dirty_cur_time = mbox->last_cur_mtime; *hdr_r = *hdr; mail_index_view_close(&view); } static int maildir_sync_quick_check(struct maildir_mailbox *mbox, const char *new_dir, const char *cur_dir, bool *new_changed_r, bool *cur_changed_r) { struct index_mailbox *ibox = &mbox->ibox; struct mail_index_header hdr; struct stat st; time_t new_mtime, cur_mtime; *new_changed_r = *cur_changed_r = FALSE; if (maildir_stat(mbox, new_dir, &st) < 0) return -1; new_mtime = st.st_mtime; if (maildir_stat(mbox, cur_dir, &st) < 0) return -1; cur_mtime = st.st_mtime; /* cur stamp is kept in index, we don't have to sync if someone else has done it and updated the index. FIXME: For now we're using sync_size field as the new/ dir's stamp. Pretty ugly.. */ maildir_sync_update_from_header(mbox, &hdr); if ((mbox->dirty_cur_time == 0 && cur_mtime != mbox->last_cur_mtime) || (new_mtime != mbox->last_new_mtime)) { /* check if the index has been updated.. */ if (mail_index_refresh(ibox->index) < 0) { mail_storage_set_index_error(ibox); return -1; } maildir_sync_update_from_header(mbox, &hdr); } /* If we're removing recent flags, always sync new/ directory if it has mails. */ if (new_mtime != mbox->last_new_mtime || ((mbox->last_dirty_flags & MAILDIR_DIRTY_NEW) != 0 && new_mtime < ioloop_time - MAILDIR_SYNC_SECS) || (!ibox->keep_recent && hdr.recent_messages_count > 0)) { *new_changed_r = TRUE; mbox->last_new_mtime = new_mtime; if (new_mtime < ioloop_time - MAILDIR_SYNC_SECS) mbox->last_dirty_flags &= ~MAILDIR_DIRTY_NEW; else mbox->last_dirty_flags |= MAILDIR_DIRTY_NEW; } if (cur_mtime != mbox->last_cur_mtime || (mbox->dirty_cur_time != 0 && ioloop_time - mbox->dirty_cur_time > MAILDIR_SYNC_SECS)) { /* cur/ changed, or delayed cur/ check */ *cur_changed_r = TRUE; mbox->last_cur_mtime = cur_mtime; if (cur_mtime < ioloop_time - MAILDIR_SYNC_SECS) { mbox->last_dirty_flags &= ~MAILDIR_DIRTY_CUR; mbox->dirty_cur_time = 0; } else { mbox->last_dirty_flags |= MAILDIR_DIRTY_CUR; mbox->dirty_cur_time = cur_mtime; } } return 0; } static int maildir_sync_context(struct maildir_sync_context *ctx, bool forced, bool sync_last_commit) { bool new_changed, cur_changed, full_rescan = FALSE; int ret; if (sync_last_commit) { new_changed = cur_changed = FALSE; } else if (!forced) { if (maildir_sync_quick_check(ctx->mbox, ctx->new_dir, ctx->cur_dir, &new_changed, &cur_changed) < 0) return -1; if (!new_changed && !cur_changed) return 1; } else { new_changed = cur_changed = TRUE; } /* Locking, locking, locking.. Wasn't maildir supposed to be lockless? We can get here either as beginning a real maildir sync, or when committing changes to maildir but a file was lost (maybe renamed). So, we're going to need two locks. One for index and one for uidlist. To avoid deadlocking do the uidlist lock always first. uidlist is needed only for figuring out UIDs for newly seen files, so theoretically we wouldn't need to lock it unless there are new files. It has a few problems though, assuming the index lock didn't already protect it (eg. in-memory indexes): 1. Just because you see a new file which doesn't exist in uidlist file, doesn't mean that the file really exists anymore, or that your readdir() lists all new files. Meaning that this is possible: A: opendir(), readdir() -> new file ... -- new files are written to the maildir -- B: opendir(), readdir() -> new file, lock uidlist, readdir() -> another new file, rewrite uidlist, unlock A: ... lock uidlist, readdir() -> nothing left, rewrite uidlist, unlock The second time running A didn't see the two new files. To handle this correctly, it must not remove the new unseen files from uidlist. This is possible to do, but adds extra complexity. 2. If another process is rename()ing files while we are readdir()ing, it's possible that readdir() never lists some files, causing Dovecot to assume they were expunged. In next sync they would show up again, but client could have already been notified of that and they would show up under new UIDs, so the damage is already done. Both of the problems can be avoided if we simply lock the uidlist before syncing and keep it until sync is finished. Typically this would happen in any case, as there is the index lock.. The second case is still a problem with external changes though, because maildir doesn't require any kind of locking. Luckily this problem rarely happens except under high amount of modifications. */ ctx->partial = !cur_changed; ret = maildir_uidlist_sync_init(ctx->mbox->uidlist, ctx->partial, &ctx->uidlist_sync_ctx); if (ret <= 0) { /* failure / timeout. if forced is TRUE, we could still go forward and check only for renamed files, but is it worth the trouble? .. */ return ret; } if (!ctx->mbox->syncing_commit) { if (maildir_sync_index_begin(ctx->mbox, ctx, &ctx->index_sync_ctx) < 0) return -1; } if (new_changed || cur_changed) { /* if we're going to check cur/ dir our current logic requires that new/ dir is checked as well. it's a good idea anyway. */ unsigned int count = 0; while ((ret = maildir_scan_dir(ctx, TRUE)) > 0) { /* rename()d at least some files, which might have caused some other files to be missed. check again (see MAILDIR_RENAME_RESCAN_COUNT). */ if (++count > MAILDIR_SCAN_DIR_MAX_COUNT) break; } if (ret < 0) return -1; if (cur_changed) { if (maildir_scan_dir(ctx, FALSE) < 0) return -1; } /* finish uidlist syncing, but keep it still locked */ maildir_uidlist_sync_finish(ctx->uidlist_sync_ctx); } if (!ctx->mbox->syncing_commit) { /* NOTE: index syncing here might cause a re-sync due to files getting lost, so this function might be called re-entrantly. FIXME: and that breaks in maildir_uidlist_sync_deinit() */ ret = maildir_sync_index(ctx->index_sync_ctx, ctx->partial); if (maildir_sync_index_finish(&ctx->index_sync_ctx, ret < 0, FALSE) < 0) return -1; if (ret < 0) return -1; if (ret == 0) full_rescan = TRUE; i_assert(maildir_uidlist_is_locked(ctx->mbox->uidlist)); } ret = maildir_uidlist_sync_deinit(&ctx->uidlist_sync_ctx); return ret < 0 ? -1 : (full_rescan ? 0 : 1); } int maildir_storage_sync_force(struct maildir_mailbox *mbox) { struct maildir_sync_context *ctx; int ret; ctx = maildir_sync_context_new(mbox); ret = maildir_sync_context(ctx, TRUE, FALSE); maildir_sync_deinit(ctx); return ret < 0 ? -1 : 0; } int maildir_sync_last_commit(struct maildir_mailbox *mbox) { struct maildir_sync_context *ctx; int ret; if (mbox->ibox.commit_log_file_seq == 0) return 0; ctx = maildir_sync_context_new(mbox); ret = maildir_sync_context(ctx, FALSE, TRUE); maildir_sync_deinit(ctx); return ret < 0 ? -1 : 0; } struct mailbox_sync_context * maildir_storage_sync_init(struct mailbox *box, enum mailbox_sync_flags flags) { struct maildir_mailbox *mbox = (struct maildir_mailbox *)box; struct maildir_sync_context *ctx; int ret = 0; if (!box->opened) index_storage_mailbox_open(&mbox->ibox); if ((flags & MAILBOX_SYNC_FLAG_FAST) == 0 || mbox->ibox.sync_last_check + MAILBOX_FULL_SYNC_INTERVAL <= ioloop_time) { mbox->ibox.sync_last_check = ioloop_time; ctx = maildir_sync_context_new(mbox); ret = maildir_sync_context(ctx, FALSE, FALSE); maildir_sync_deinit(ctx); i_assert(!maildir_uidlist_is_locked(mbox->uidlist) || mbox->ibox.keep_locked); if (ret == 0) { /* lost some files from new/, see if thery're in cur/ */ ret = maildir_storage_sync_force(mbox); } } return index_mailbox_sync_init(box, flags, ret < 0); } int maildir_sync_is_synced(struct maildir_mailbox *mbox) { const char *new_dir, *cur_dir; bool new_changed, cur_changed; int ret; t_push(); new_dir = t_strconcat(mbox->path, "/new", NULL); cur_dir = t_strconcat(mbox->path, "/cur", NULL); ret = maildir_sync_quick_check(mbox, new_dir, cur_dir, &new_changed, &cur_changed); t_pop(); return ret < 0 ? -1 : (!new_changed && !cur_changed); }