Date | Commit message (Collapse) |
|
This should help us deal with MH sequence number packing and
invalidating mail_sync.sqlite3.
|
|
This was a single-caller sub since 47d4e53734820b4e
(lei_mail_sync: rely on flock(2), avoid IPC, 2021-09-18)
and unlikely to be used further, so inline it and save
a few KB of memory.
|
|
Enqueuing "note-event" requests from the DS event loop must
not wait on workers being able to drain the queue quickly
enough. Thus we make the SOCK_SEQPACKET writes nonblocking
and rely on the lei-daemon event loop to enqueue writes.
This is a unique problem for "note-event" since it reuses
workers in between commands, while most lei commands currently
fork off new workers.
|
|
This change prevents lingering shard and git-fast-import
processes from remaining after interrupted "lei import" (and
similar). It also reduces the likelyhood of data-loss in case
of subsequent abnormal termination of the daemon.
I think this is the least surprising way to handle users
prematurely aborting imports or other similar operations which
write to lei/store and will result in reduced bandwidth waste
for users with intermittent connections. This is because the
lei/store processes may be shared by parallel "lei import"
callers, and commits done by any "lei import" caller will
inevitably trigger writes for all of them.
|
|
When a file goes away, try to make sure we don't waste
time trying to access or store it.
|
|
No need to pass extra arrayref args, here.
|
|
We don't want to lose the failure message in case note-event
fails.
|
|
We need to send the socket over to lei/store and wait for the
kernel to drop the socket refcount down to zero before
script/lei can exit.
This is not a new bug and only caused very sporadic test
failures. I only noticed it while simplifying IPC stuff.
|
|
Simplify our APIs and force dwaitpid() to work in async mode for
all lei workers. This avoids having lingering zombies for
parallel searches if one worker finishes soon before another.
The old distinction between "old" and "new" workers was
needlessly complex, error-prone, and embarrasingly bad.
We also never handled v2:// writers properly before on
Ctrl-C/Ctrl-Z (SIGINT/SIGTSTP), so add them to @WQ_KEYS
to ensure they get handled by $lei when appropropriate.
|
|
Instead of passing the prefix section and key separately, pass
them together as is commonly done with git-config(1) usage as
well as our ->get_all API. This inconsistency in the get_1 API
is a needless footgun and confused me a bit while working on
"lei up" the other week.
|
|
The note-event worker may see changes before a Xapian shard
commit happens, meaning keyword lookups fail as a result.
Just emit the request to the lei/store worker since it's a
fairly cheap operation at this point.
We'll try harder to look for kw changes, too, since
deduplication changes may lead to multiple docids being
resolved for a single message.
|
|
Redundant code is noise and therefore confusing :<
|
|
We can set opt->{quiet} for (internal) 'note-event' command
to quiet ->qerr, since we use ->qerr everywhere else. And
we'll just die() instead of setting a ->{fail} message, since
eval + die are more inline with the rest of our Perl code.
|
|
"All" my CPUs is only 4, but it's probably ridiculous for
somebody with a 16-core system to have 16 processes for
accessing SQLite DBs.
We do the same thing in Pmdir for parallel Maildir access
(and V2Writable).
|
|
This has several advantages:
* no need to use ipc.lock to protect a pipe for non-atomic writes
* ability to pass FDs. In another commit, this will let us
simplify lei->sto_done_request and pass newly-created
sockets to lei/store directly.
disadvantages:
- an extra pipe is required for rare messages over several
hundred KB, this is probably a non-issue, though
The performance delta is unknown, but I expect shards
(which remain pipes) to be the primary bottleneck IPC-wise
for lei/store.
|
|
A common pattern we use is to arm a timer once and prevent
it from being armed until it fires. We'll be using it more
to do polling for saved searches and imports.
|
|
Since 44917fdd24a8bec1 ("lei_mail_sync: do not use transactions"),
relying on lei/store to serialize access was a pointless endeavor.
Rely on flock(2) to serialize multiple writers since (in my
experience) it's the easiest way to deal with parallel writers
when using SQLite. This allows us to simplify existing callers
while speeding up 'lei refresh-mail-sync --all=local' by 5% or
so.
|
|
This works with existing inotify/EVFILT_VNODE functionality to
propagate changes made from one Maildir to another Maildir.
I chose the lei/store worker process to handle this since
propagating changes back into lei-daemon on a massive scale
could lead to dead-locking while both processes are attempting
to write to each other. Eliminating IPC overhead is a nice
side effect, but could hurt performance if Maildirs are slow.
The code for "lei export-kw" is significantly revamped to match
the new code used in the "lei/store" daemon. It should be more
correct w.r.t. corner-cases and stale entries, but perhaps
better tests need to be written.
squashed:
t/lei-auto-watch: increase delay for FreeBSD kevent
My FreeBSD VM seems to need longer for this test than inotify
under Linux, likely because the kevent support code needs to be
more complicated.
|
|
This allows client sockets to wait for "done" commits to
lei/store while the daemon reacts asynchronously. The goal
of this change is to keep the script/lei client alive until
lei/store commits changes to the filesystem, but without
blocking the lei-daemon event loop. It depends on Perl
refcounting to close the socket.
This change also highlighted our over-use of "done" requests to
lei/store processes, which is now corrected so we only issue it
on collective socket EOF rather than upon reaping every single
worker.
This also fixes "lei forget-mail-sync" when it is the initial
command.
This took several iterations and much debugging to arrive at the
current implementation:
1. The initial iteration of this change utilized socket passing
from lei-daemon to lei/store, which necessitated switching
from faster pipes to slower Unix sockets.
2. The second iteration switched to registering notification sockets
independently of "done" requests, but that could lead to early
wakeups when "done" was requested by other workers. This
appeared to work most of the time, but suffered races under
high load which were difficult to track down.
Finally, this iteration passes the stringified socket GLOB ref
to lei/store which is echoed back to lei-daemon upon completion
of that particular "done" request.
|
|
This allows MUA-made flag changes to Maildirs to be instantly
read and acknowledged for future search results.
In the future, it may be used to speed up --augment and
--import-before (the default) with with "lei q".
|
|
This makes behavior less surprising on restarts as we no longer
lose state on restarts, so there's no need to manually run "lei
add-watch" to re-enable watches. This also allows us to
transparently handle changes if somebody edits the lei config
file directly or via git-config(1).
|
|
This allows lei to automatically note keyword (message flag)
changes made to a Maildir and propagate it into lei/store:
lei add-watch --state=tag-ro /path/to/Maildir
This doesn't persist across restarts, yet. In the future,
it will be applied automatically to "lei q" output Maildirs
by default (with an option to disable it).
State values of tag-rw, index-<ro|rw>, import-<ro|rw> will all
be supported for Maildir.
This represents a fairly major internal change that's fairly
intrusive, but the whole daemon-oriented design was to
facilitate being able to automatically monitor (and propagate)
Maildir/IMAP flag changes.
|