Date | Commit message (Collapse) |
|
Testing with perl-5.16.3-294.el7_6 RPM package on RHEL/CentOS 7,
the Deflater middleware triggers a leak when used in conjunction
with our push-based responses from PublicInbox::Qspawn.
I could not find another solution to workaround the memory leak
in this case, and I could not find a specific leak fix in
the perl5180delta manpage[1] which looked like it would
solve our problem.
Attempting to workaround the issue proved futile. Using
internal Deflater-specific keys to prevent deflating in
GitHTTPBackend and Qspawn did not solve the problem:
$env->{"plack.skip-deflater"} = 1;
$env->{"psgix.no-compress"} = 1;
Nor did forcing an invalid encoding via "git fetch":
git -c http.extraheader=Accept-Encoding:gzap fetch
So this appears to be a problem with Plack::Util::response_cb
somewhere.
This does NOT appear to be a problem with ref() leaking as in
DS::next_tick[2], since I couldn't find where
Plack::Middleware::Deflater or Plack::Util::response_cb would be
calling ref() on a blessed reference to trigger a leak.
Also, oddly enough, the ref() use for backwards compatibility at
the top of PublicInbox::GitHTTPBackend::serve does NOT seem to
trigger a leak on 5.16.3 due to [2]:
# XXX compatibility... ugh, can we stop supporting this?
$git = PublicInbox::Git->new($git) unless ref($git);
[1] https://perldoc.perl.org/perl5180delta.html
[2] https://rt.perl.org/Public/Bug/Display.html?id=114340
|
|
The perl-5.16.3-294.el7_6 RPM package on RHEL/CentOS 7 is
affected by a memory leak in Perl when calling `ref' on
blessed references. This resulted in a very slow leak that
manifests more quickly with a nonstop "git fetch" loop.
Use Scalar::Util::blessed to work around the issue.
Tested overnight on a CentOS 7 VM.
cf. https://rt.perl.org/Public/Bug/Display.html?id=114340
|
|
We support "-env" to clear the environment with spawn(),
which causes test failures but no runtime failures
(since "-env" isn't used anywhere in our real code)
Reported-and-tested-by: Alyssa Ross <hi@alyssa.is>
|
|
It was never used, and will not be needed.
|
|
Inline::C seems alright, so we might use it more since it still
allows end users to quickly make changes. Our performance on
rotational disks is also terrible, and could be improved...
|
|
We were emitting the same "<id>mailto:name@domain</id>" tag
for every feed (but not per-feed entry). This could cause
feed readers to mistake the top (news.atom) feed for other
feeds (search results, or per-thread feeds).
This is technically a breaking change for people relying on
per-thread or per-query feeds, but the only alternative is
to remain broken for anybody trying to follow multiple feeds
off the same inbox.
|
|
We need to handle arbitrary integers and case-insensitive
variations of human words to match git-config(1) behavior,
since that's what users would expect given we use config
files parseable by git-config(1).
|
|
The v1.2.0 is a work-in-progress, while the others are copied
out of our mail archives.
Eventually, a NEWS file will be generated from these emails and
distributed in the release tarball. There'll also be an Atom
feed for the website reusing our feed generation code.
|
|
Actually do the redirect properly
|
|
We don't want the stdin from the test runner to accidentally
cause this test to fail.
|
|
Only removing $http->{env} is needed to prevent circular
references. $env->{'psgix.io'} does not need to be deleted
since $env will no longer have any references to it when
->close returns.
|
|
And explain why we need to do that delete in a comment.
|
|
Naming $start_cb consistently helps avoid confusing new readers,
and some comments will help with understanding flow
|
|
All of these circular references are designed to clear
themselves, but these will make actual errors from Devel::Cycle
easier-to-spot.
The circular reference in the limiter {run_queue} is not a real
problem, but we can avoid storing the circular reference until
we actually need to spawn the child, reducing the size of the
Qspawn object while it's in the queue, slightly.
We also do not need to have redundant checks to spawn new
processes, we should only spawn new processes when they're
->start-ed or after waitpid reaps them.
|
|
Generic PSGI servers have $env->{'psgi.errors'}, too,
so ensure they can log errors.
|
|
We don't use the return value in real code since we do waitpid
asynchronously, now. So simplify our runtime code at the cost
of making our test slighly more complex.
|
|
It's a bit of an esoteric option, but maybe somebody out
there can find it useful.
|
|
We don't need to hold onto the subprocess environ and
redirects/options for popen_rd after spawning the child process.
I do not expect this to fix problem of leaking unlinked regular
file descriptors (which I still can't reproduce), and it
definitely does not fix the problem of leaking pipe descriptors
(which I also can't reproduce).
This will save an FD sooner on non-public-inbox-httpd servers
which give a non-FD $env->{'psgi.input'}, however
Regardless, it's good to free up memory resources in our own
process ASAP we're done using them.
|
|
EINTR should not happen when using non-blocking sockets like we
do in our daemons, but maybe some OSes allow it to happen and
edge-triggered notifications won't notify us again.
So always retry immediately on EINTR without relying on kqueue
or epoll to notify us, and log any other unrecoverable errors
which may happen while we're at it.
|
|
We already import `which' for lsof(8), so we might as well
use it to detect curl(1), too.
|
|
NNTPS and STARTTLS seems to be working for several months
without incident on news.public-inbox.org, so consider it a
success and maybe others can try using it.
HTTPS technically works, too, but isn't documented at
the moment since I can't recommend production deployments
without varnish protecting it.
|
|
-W0 (no workers) should not create any pipes on its own,
and we shouldn't have any deleted FDs if no clients are
connected.
This can find if leaks which may be triggered by PublicInbox::HTTP
(and not Qspawn or GitHTTPBackend).
|
|
We rely on DS to do waitpid with WNOHANG, now, and the non-DS
code path won't use WNOHANG.
|
|
Rename the {cleanup} field to {end}, since it's similar
to END {} and is consistent with the variable in Qspawn.pm
And document how certain subs get called, since we have
many subs named "new" and "close".
|
|
REMOTE_HOST is not set by us (it is the reverse DNS name) of
REMOTE_ADDR, and there's few better ways to kill HTTP server
performance than to use standard name resolution APIs like
getnameinfo(3).
|
|
Might as well share some code for temporary file creation
|
|
Although we always unlink temporary files, give them a
meaningful name so that we can we can still make sense
of the pre-unlink name when using lsof(8) or similar
tools on Linux.
|
|
I didn't know PerlIO::scalar existed until a few months ago,
but it's been distributed with Perl since 5.8 and doesn't
seem to be split out into it's own package on any distro.
|
|
For whatever reason, inbox directories can go missing
temporarily or permanently. Tell the admin about them
and continue on our way.
|
|
It's possible for Qspawn callers to be deferred, in which case
we must ensure we don't cause the temporary file used for
stdin to become unref-ed and closed.
This can be a problem when we exceed the default Qspawn
limiter of 32 concurrent processes for "git update-index".
|
|
Users of socket activation don't need it, and hopefully other
init systems support it, too.
|
|
Quiets down pod2man complaining
|
|
|
|
It was never documented, before.
|
|
IO::Socket::INET->new is rather verbose with the options hash,
extract it into a standalone sub
|
|
We don't want to get hung into a state where we see "\n" via
index(), yet cannot consume rbuf in the while loop. So tweak
the regexp to ensure we always consume rbuf.
I suspect this is what causes occasional 100% CPU usage of
-nntpd, but reproducing it's been difficult..
|
|
Since Net::NNTP::listgroup doesn't support the range parameter,
I had to test this manually and noticed extra CRLF were emitted.
|
|
We don't need to keep an empty buffer around in the common case
when a client is sending us completely inflatable requests and
we're able to read them in one go.
This only seems to save about 2M with 10K NNTPS clients using
COMPRESS, so it's not a huge win, but better than nothing.
|
|
It never ends...
|
|
RFC3977 6.1.2.2 LISTGROUP allows a [range] arg after [group],
and supporting it allows NNTP support in neomutt to work again.
Tested with NeoMutt 20170113 (1.7.2) on Debian stretch
(oldstable)
|
|
RFC3977 8.4.2 mandates the order of non-standard headers
to be after the first seven standard headers/metadata;
so "Xref:" must appear after "Lines:"|":lines".
Additionally, non-required header names must be followed
by ":full".
Cc: Jonathan Corbet <corbet@lwn.net>
Reported-by: Urs Janßen
<E1hmKBw-0008Bq-8t@akw>
|
|
We won't need further layering after enabling compression. So
be explicit about which sub we're calling when we hit ->do_read
from NNTP and eliminate the need for the comment.
|
|
We need to ensure further timers can be registered if there's
currently no idle clients.
|
|
In HTTP.pm, we can use the same technique NNTP.pm uses with
long_response with the $long_cb callback and avoid storing
$pull in the per-client structure at all. We can also reuse
the same logic to push the callback into wbuf from NNTP.
This does NOT introduce a new circular reference, but documents
it more clearly.
|
|
We're using Qspawn, now
|
|
Relying on "use" to import during BEGIN means we get to take
advantage of prototype checking of function args during the rest
of the compilation phase.
|
|
*glob notation isn't always necessary, and there's
no need to disable 'once' warnings, this way.
|
|
No point in uglifying our code since we need the POSIX
module in many places, anyways.
|
|
* origin/nntp-compress:
nntp: improve error reporting for COMPRESS
nntp: reduce memory overhead of zlib
nntp: support COMPRESS DEFLATE per RFC 8054
nntp: move LINE_MAX constant to the top
nntp: use msg_more as a method
|
|
While we're usually not stuck waiting on waitpid after
seeing a pipe EOF or even triggering SIGPIPE in the process
(e.g. git-http-backend) we're reading from, it MAY happen
and we should be careful to never hang the daemon process
on waitpid calls.
v2: use "eq" for string comparison against 'DEFAULT'
|