All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jan Kara <jack@suse.cz>
To: mathieu lacage <mathieu.lacage@alcmeon.com>
Cc: jack@suse.cz,
	Olivier Chapelliere <olivier.chapelliere@alcmeon.com>,
	linux-fsdevel@vger.kernel.org
Subject: Re: stuck in inotify_release
Date: Tue, 14 May 2019 17:44:05 +0200	[thread overview]
Message-ID: <20190514154405.GA30401@quack2.suse.cz> (raw)
In-Reply-To: <CAC8Mkjy=igiQatSVXNXphjyzGn2faZ75XZZGANWOtt3hvwk8DA@mail.gmail.com>

Hi!

On Tue 14-05-19 16:35:29, mathieu lacage wrote:
> We are going to setup a new ubuntu 16.04 server, rebuild a vanilla 5.0
> kernel on that and run a fraction of our production workload on that. Is
> this ok for you ? If so, I will let you know as soon as we observe the
> problem on this server again.

Yes, that should rule out any Ubuntu specific problems thanks!

								Honza

> 
> Mathieu
> 
> Le mar. 14 mai 2019 à 15:22, Olivier Chapelliere <
> olivier.chapelliere@alcmeon.com> a écrit :
> 
> > ---------- Forwarded message ---------
> > From: Jan Kara <jack@suse.cz>
> > Date: Tue, May 14, 2019 at 11:25 AM
> > Subject: Re: stuck in inotify_release
> > To: Olivier Chapelliere <olivier.chapelliere@alcmeon.com>
> > Cc: Jan Kara <jack@suse.cz>, <linux-fsdevel@vger.kernel.org>
> >
> >
> > Hello!
> >
> > On Mon 06-05-19 20:54:24, Olivier Chapelliere wrote:
> > > It finally took a month to happen again : python processes watching a
> > > directory are stuck in inotify_release.
> > > I ran the sysrq commands as you requested and attached the result.
> >
> > Thanks. I was looking into these traces but the situation is the same as
> > before. Everyone is blocked waiting for inotify group to shut down. That is
> > blocked waiting for worker to finish destroying notification marks and the
> > worker is blocked in synchronize_srcu() waiting for SRCU grace period to
> > end. Now I didn't find any process that would be holding the SRCU lock so
> > it seems that someone exited the SRCU locked section without releasing the
> > lock. I've checked 4.15 your Ubuntu kernel is based on and I don't see how
> > that would be possible. It it possible though, that the problem is
> > introduced by some Ubuntu specific backports. Would it be possible for you
> > to run some vanilla kernel (i.e., without Ubuntu modifications)?
> >
> >                                                                 Honza
> >
> > > On Thu, Mar 28, 2019 at 10:52 AM Jan Kara <jack@suse.cz> wrote:
> > > >
> > > > Hello,
> > > >
> > > > On Thu 28-03-19 09:26:45, Olivier Chapelliere wrote:
> > > > > According to what I read on internet you seem to be the right person
> > to get
> > > > > in touch with when one has problems with inotify.
> > > >
> > > > Yes, there's also linux-fsdevel@vger.kernel.org mailing list which we
> > use
> > > > (added to CC).
> > > >
> > > > > We are monitoring several directories in python processes through
> > inotify.
> > > > > But after few days all processes are stuck in a call to
> > inotify_release.
> > > > > Once I detected the problem, I dumped info to dmesg with
> > sysrq-trigger
> > > > > (dmesg content attached):
> > > > > echo w > /proc/sysrq-trigger
> > > >
> > > > Looking through the stack traces, all of them wait in fput() ->
> > > > inotify_release() -> ... -> fsnotify_wait_marks_destroyed() ->
> > > > flush_delayed_work(&reaper_work). So they wait for worker process to
> > > > destroy all marks for the group. However that worker (kworker/u8:4) is
> > > > stuck in:
> > > >
> > > > fsnotify_mark_destroy_workfn() -> synchronize_srcu(&fsnotify_mark_srcu)
> > > >
> > > > So the question is who is holding fsnotify_mark_srcu so that SRCU
> > cannot
> > > > declare new grace period. I don't see any such process among the
> > processes
> > > > you've shown in the dump (but it should be there) so it's a bit of a
> > > > mystery.
> > > >
> > > > > Our production env is ubuntu 18.04 kernel 4.15 fs ext4
> > > > > This problem appears on a weekly basis so I will be able to run
> > additional
> > > > > commands to track down the issue if needed.
> > > >
> > > > So when this happens again, try grabbing output of sysrq-l and sysrq-t
> > if
> > > > we can find the task holding fsnotify_mark_srcu.
> > > >
> > > >                                                                 Honza
> > > > --
> > > > Jan Kara <jack@suse.com>
> > > > SUSE Labs, CR
> > >
> > >
> > >
> > > --
> > > Olivier Chapelliere
> >
> >
> > --
> > Jan Kara <jack@suse.com>
> > SUSE Labs, CR
> >
> >
> > --
> > Olivier Chapelliere
> >
> 
> 
> -- 
> Mathieu Lacage <mathieu.lacage@alcmeon.com>
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR

      parent reply	other threads:[~2019-05-14 15:44 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <CANp+0hhbsegocrx-MK0DS=Qx4DfivB27nSKHrukiFAY6x6cJQA@mail.gmail.com>
2019-03-28  9:52 ` stuck in inotify_release Jan Kara
2019-05-06 18:54   ` Olivier Chapelliere
2019-05-14  6:22     ` Olivier Chapelliere
2019-05-14  9:25     ` Jan Kara
     [not found]       ` <CANp+0hiZt=oEWMUqRC-pv9=8JnvSyPcpDCf+O5whth1C_q0jNA@mail.gmail.com>
     [not found]         ` <CAC8Mkjy=igiQatSVXNXphjyzGn2faZ75XZZGANWOtt3hvwk8DA@mail.gmail.com>
2019-05-14 15:44           ` Jan Kara [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20190514154405.GA30401@quack2.suse.cz \
    --to=jack@suse.cz \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=mathieu.lacage@alcmeon.com \
    --cc=olivier.chapelliere@alcmeon.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.