All of lore.kernel.org
 help / color / mirror / Atom feed
From: Kevin Decherf <kevin@kdecherf.com>
To: Sam Lang <sam.lang@inktank.com>
Cc: "ceph-devel@vger.kernel.org" <ceph-devel@vger.kernel.org>,
	support@clever-cloud.com
Subject: Re: Crash and strange things on MDS
Date: Mon, 11 Feb 2013 19:54:24 +0100	[thread overview]
Message-ID: <20130211185424.GA27669@kdecherf.com> (raw)
In-Reply-To: <CAKMAVE_J4GOA_yUF5ue-y+_iFVhbvCqaGPvBOfgtEuO7CzRU6g@mail.gmail.com>

On Mon, Feb 11, 2013 at 11:00:15AM -0600, Sam Lang wrote:
> Hi Kevin, sorry for the delayed response.
> This looks like the mds cache is thrashing quite a bit, and with
> multiple MDSs the tree partitioning is causing those estale messages.
> In your case, you should probably run with just a single active mds (I
> assume all three MDSs are active, but ceph -s will tell you for sure),
> and the others as standby.  I don't think you'll be able to do that
> without starting over though.

Hi Sam,

I know that MDS clustering is a bit buggy so I have only one active MDS
on this cluster.

Here is the output of ceph -s:

   ~ # ceph -s
      health HEALTH_OK
      monmap e1: 3 mons at {a=x:6789/0,b=y:6789/0,c=z:6789/0}, election epoch 48, quorum 0,1,2 a,b,c
      osdmap e79: 27 osds: 27 up, 27 in
       pgmap v895343: 5376 pgs: 5376 active+clean; 18987 MB data, 103 GB used, 21918 GB / 23201 GB avail
      mdsmap e73: 1/1/1 up {0=b=up:active}, 2 up:standby


> Also, you might want to increase the size of the mds cache if you have
> enough memory on that machine.  mds cache size defaults to 100k, you
> might increase it to 300k and see if you get the same problems.

I have 24GB of memory for each MDS, I will try to increase this value.
Thanks for advice.

> Do you have debug logging enabled when you see this crash?  Can you
> compress that mds log and post it somewhere or email it to me?

Yes, I have 34GB of raw logs (for this issue) but I have no debug log
of the beginning of the storm itself. I will upload a compressed
archive.


Furthermore, I observe another strange thing more or less related to the
storms.

During a rsync command to write ~20G of data on Ceph and during (and
after) the storm, one OSD sends a lot of data to the active MDS
(400Mbps peak each 6 seconds). After a quick check, I found that when I
stop osd.23, osd.14 stops its peaks.

I will forward a copy of the debug enabled log of osd14.

The only significant difference between osd.23 and others is the list of
hb_in where osd.14 is missing (but I think it's unrelated).

   ~ # ceph pg dump
   osdstat  kbused   kbavail  kb hb in hb out
   0  4016228  851255948   901042464   [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26]  []
   1  4108748  851163428   901042464   [0,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,18,19,20,21,22,23,24,25,26]  []
   2  4276584  850995592   901042464   [0,1,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26]  []
   3  3997368  851274808   901042464   [0,1,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26] []
   4  4358212  850913964   901042464   [0,1,2,3,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26]  []
   5  4039112  851233064   901042464   [0,1,2,3,4,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26]  []
   6  3971568  851300608   901042464   [0,1,2,3,4,5,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26]  []
   7  3942556  851329620   901042464   [0,1,2,3,4,5,6,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26]  []
   8  4275584  850996592   901042464   [0,1,2,3,4,5,6,7,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26]  []
   9  4279308  850992868   901042464   [0,1,2,3,4,5,6,7,8,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26]  []
   10 3728136  851544040   901042464   [0,1,2,3,4,5,6,7,8,9,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26]   []
   11 3934096  851338080   901042464   [0,1,2,3,4,5,6,7,8,9,10,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26]   []
   12 3991600  851280576   901042464   [0,1,2,3,4,5,6,7,8,9,10,11,13,14,15,16,17,18,19,20,21,22,23,24,25,26]   []
   13 4211228  851060948   901042464   [0,1,2,3,4,5,6,7,8,9,10,11,12,14,15,16,17,18,19,20,21,22,23,24,25,26]   []
   14 4169476  851102700   901042464   [0,1,2,3,4,5,6,7,8,9,10,11,12,13,15,16,17,18,19,20,21,22,23,24,25,26]   []
   15 4385584  850886592   901042464   [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,16,17,18,19,20,21,22,23,24,25,26]   []
   16 3761176  851511000   901042464   [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,17,18,19,20,21,22,23,24,25,26]   []
   17 3646096  851626080   901042464   [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,18,19,20,21,22,23,24,25,26]   []
   18 4119448  851152728   901042464   [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,19,20,21,22,23,24,25,26]   []
   19 4592992  850679184   901042464   [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,20,21,22,23,24,25,26]   []
   20 3740840  851531336   901042464   [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,21,22,23,24,25,26]   []
   21 4363552  850908624   901042464   [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,22,23,24,25,26]   []
   22 3831420  851440756   901042464   [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,23,24,25,26]   []
   23 3681648  851590528   901042464   [0,1,2,3,4,5,6,7,8,9,10,11,12,13,15,16,17,18,19,20,21,22,24,25,26]   []
   24 3946192  851325984   901042464   [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,25,26]   []
   25 3954360  851317816   901042464   [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,26]   []
   26 3775532  851496644   901042464   [0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25]   []
    sum  109098644   22983250108 24328146528

Cheers,
-- 
Kevin Decherf - @Kdecherf
GPG C610 FE73 E706 F968 612B E4B2 108A BD75 A81E 6E2F
http://kdecherf.com

  reply	other threads:[~2013-02-11 18:54 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-02-04 18:01 Crash and strange things on MDS Kevin Decherf
2013-02-11 13:05 ` Kevin Decherf
2013-02-11 17:00   ` Sam Lang
2013-02-11 18:54     ` Kevin Decherf [this message]
2013-02-11 20:25       ` Gregory Farnum
2013-02-11 22:24         ` Kevin Decherf
2013-02-11 22:47           ` Gregory Farnum
2013-02-11 23:33             ` Kevin Decherf
2013-02-13 11:47         ` Kevin Decherf
2013-02-13 18:19           ` Gregory Farnum
2013-02-16  1:02             ` Kevin Decherf
2013-02-16 17:36               ` Sam Lang
2013-02-16 18:24                 ` Kevin Decherf
2013-02-19 18:15                   ` Gregory Farnum
2013-02-20  1:00                     ` Kevin Decherf
2013-02-20  1:09                       ` Gregory Farnum
2013-02-26 17:57                         ` Kevin Decherf
2013-02-26 18:10                           ` Gregory Farnum
2013-02-26 19:58                             ` Kevin Decherf
2013-02-26 20:26                               ` Gregory Farnum
2013-02-26 21:57                                 ` Kevin Decherf
2013-02-26 21:58                                   ` Gregory Farnum
2013-02-27  0:03                                     ` Yan, Zheng 
2013-02-27  0:14                                       ` Sage Weil
     [not found]                                     ` <20130227004923.GQ16091@kdecherf.com>
     [not found]                                       ` <CAPYLRzhbygkA9=DkVr474Nw8AOC2hAFG-1D6uS4WyfR=kUBXWQ@mail.gmail.com>
     [not found]                                         ` <20130308232943.GA2197@kdecherf.com>
     [not found]                                           ` <20130308232943.GA2197-fShu9kyPgSlWk0Htik3J/w@public.gmane.org>
2013-03-15 20:32                                             ` Greg Farnum
     [not found]                                               ` <ECAA10260D284057A52D78127F8634A8-4GqslpFJ+cxBDgjK7y7TUQ@public.gmane.org>
2013-03-15 22:40                                                 ` Marc-Antoine Perennou
2013-03-15 22:53                                                   ` Greg Farnum

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=20130211185424.GA27669@kdecherf.com \
    --to=kevin@kdecherf.com \
    --cc=ceph-devel@vger.kernel.org \
    --cc=sam.lang@inktank.com \
    --cc=support@clever-cloud.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.