linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Marc MERLIN <marc@merlins.org>
To: "Andrew E. Mileski" <andrewm@isoar.ca>
Cc: linux-btrfs@vger.kernel.org
Subject: Re: Is btrfs related to OOM death problems on my 8GB server with both 3.15.1 and 3.14?
Date: Sun, 6 Jul 2014 07:58:15 -0700	[thread overview]
Message-ID: <20140706145815.GD15009@merlins.org> (raw)
In-Reply-To: <20140705144318.GT26932@merlins.org>

On Sat, Jul 05, 2014 at 07:43:18AM -0700, Marc MERLIN wrote:
> On Sat, Jul 05, 2014 at 09:47:09AM -0400, Andrew E. Mileski wrote:
> > On 2014-07-03 9:19 PM, Marc MERLIN wrote:
> > >I upgraded my server from 3.14 to 3.15.1 last week, and since then it's been
> > >running out of memory and deadlocking (panic= doesn't even work).
> > >I downgraded back to 3.14, but I already had the problem once since then.
> > 
> > I didn't see any mention of the btrfs utility version in this thread
> > (I may be blind though).
> > 
> > My server was suffering from frequent panics upon scrub / defrag /
> > balance, until I updated the btrfs utility.  That resolved all my
> > issues.
> 
> Really? The userland tool should only send ioctls to the kernel, I
> really can't see how it would cause the kernel code to panic or not.
> 
> gargamel:~# btrfs --version
> Btrfs v3.14.1
> which is the latest in debian unstable.
> 
> As an update, after 1.7 days of scrubbing, the system has started
> getting sluggish, I'm getting synchronization problems/crashes in some of
> my tools that talk to serial ports (likely due to mini deadlocks in the
> kernel), and I'm now getting a few btrfs hangs.

Predictably, it died yesterday afternoon after going into memory death
(it was answering pings, but userspace was dead, and even sysrq-o did
not respond, I had to power cycle the power outlet).

This happened just before my 3rd scrub finished, so I'm now 2 out of 2:
running scrub on my 3 filesystems kills the system half way through the
3rd scrub.

This is the last memory log that reached the disk:
http://marc.merlins.org/tmp/btrfs-oom2.txt

Do those logs point to any possible culprit, or a kernel memory leak
cannot be pointed to its source because the kernel loses track of who
requested the memory that leaked?


Excerpt here:
Sat Jul  5 14:25:04 PDT 2014
             total       used       free     shared    buffers     cached
Mem:       7894792    7712384     182408          0         28     227480
-/+ buffers/cache:    7484876     409916
Swap:     15616764     463732   15153032

Userspace is using 345MB according to ps

Sat Jul  5 14:25:04 PDT 2014
MemTotal:        7894792 kB
MemFree:          184556 kB
MemAvailable:     269568 kB
Buffers:              28 kB
Cached:           228164 kB
SwapCached:        18296 kB
Active:           178196 kB
Inactive:         187016 kB
Active(anon):      70068 kB
Inactive(anon):    71100 kB
Active(file):     108128 kB
Inactive(file):   115916 kB
Unevictable:        5624 kB
Mlocked:            5624 kB
SwapTotal:      15616764 kB
SwapFree:       15152768 kB
Dirty:             17716 kB
Writeback:           516 kB
AnonPages:        140588 kB
Mapped:            21940 kB
Shmem:               688 kB
Slab:             181708 kB
SReclaimable:      59808 kB
SUnreclaim:       121900 kB
KernelStack:        4728 kB
PageTables:         8480 kB
NFS_Unstable:          0 kB
Bounce:                0 kB
WritebackTmp:          0 kB
CommitLimit:    19564160 kB
Committed_AS:    1633204 kB
VmallocTotal:   34359738367 kB
VmallocUsed:      358996 kB
VmallocChunk:   34359281468 kB
HardwareCorrupted:     0 kB
AnonHugePages:         0 kB
HugePages_Total:       0
HugePages_Free:        0
HugePages_Rsvd:        0
HugePages_Surp:        0
Hugepagesize:       2048 kB
DirectMap4k:      144920 kB
DirectMap2M:     7942144 kB

Thanks,
Marc
-- 
"A mouse is a device used to point at the xterm you want to type in" - A.S.R.
Microsoft is to operating systems ....
                                      .... what McDonalds is to gourmet cooking
Home page: http://marc.merlins.org/                         | PGP 1024R/763BE901

  parent reply	other threads:[~2014-07-06 14:58 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-07-04  1:19 Is btrfs related to OOM death problems on my 8GB server with both 3.15.1 and 3.14? Marc MERLIN
2014-07-04  4:33 ` Russell Coker
2014-07-04  6:04   ` Marc MERLIN
2014-07-04  6:23   ` Satoru Takeuchi
2014-07-04 14:24     ` Marc MERLIN
2014-07-04 14:45       ` Russell Coker
2014-07-04 15:07         ` Marc MERLIN
2014-07-04 22:13       ` Duncan
2014-07-05 13:47 ` Andrew E. Mileski
2014-07-05 14:43   ` Marc MERLIN
2014-07-05 15:17     ` Andrew E. Mileski
2014-07-06 14:58     ` Marc MERLIN [this message]
2014-07-13 14:29       ` btrfs is " Marc MERLIN
2014-07-13 15:37         ` Marc MERLIN
2014-07-13 15:45           ` btrfs quotas " Marc MERLIN
2014-07-14  1:36             ` Qu Wenruo
2014-07-14  2:43               ` Marc MERLIN
2014-07-14  1:24         ` btrfs is " Qu Wenruo
2014-07-16  0:36           ` Jérôme Poulin
2014-07-16 15:55           ` Marc MERLIN
2014-07-17  2:22             ` Marc MERLIN
2014-07-16  0:45       ` Is btrfs " Jérôme Poulin
2014-07-05 14:27 ` Andrew E. Mileski

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=20140706145815.GD15009@merlins.org \
    --to=marc@merlins.org \
    --cc=andrewm@isoar.ca \
    --cc=linux-btrfs@vger.kernel.org \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).