Re: mds crash loop

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, Nov 12, 2019 at 6:18 PM Karsten Nielsen <karsten@xxxxxxxxxx> wrote:
>
> -----Original message-----
> From:   Karsten Nielsen <karsten@xxxxxxxxxx>
> Sent:   Tue 12-11-2019 10:30
> Subject:         Re: mds crash loop
> To:     Yan, Zheng <ukernel@xxxxxxxxx>;
> CC:     ceph-users@xxxxxxx;
> > -----Original message-----
> > From: Yan, Zheng <ukernel@xxxxxxxxx>
> > Sent: Mon 11-11-2019 15:09
> > Subject:      Re:  Re: mds crash loop
> > To:   Karsten Nielsen <karsten@xxxxxxxxxx>;
> > CC:   ceph-users@xxxxxxx;
> > > On Mon, Nov 11, 2019 at 5:09 PM Karsten Nielsen <karsten@xxxxxxxxxx> wrote:
> > > >
> > > > I started a job that moved some files around in the cephfs cluster that
> > > resulted in the mds to go back into the crash loop.
> > > > Logs are here:
> > > > http://s3.foo-bar.dk/mds-dumps/mds.log-20191111
> > > >
> > > > Any help would be appriciated.
> > > >
> > >
> > > looks like snaptable is corrupted.
> > >
> > > nautilus version (14.2.2) of ‘cephfs-data-scan scan_links’  can fix
> > > snaptable. hopefully it will fix your issue.
> > >
> > > you don't need to upgrade whole cluster. Just install nautilus in a
> > > temp machine or compile ceph from source.
> >
> > I did run the command that you suggested, it did not unfortunately fix the
> > problem.
> >
> > http://s3.foo-bar.dk/mds-dumps/mds.log-20191112
> >
>
>
> The output from the command is this:
>
> sudo docker exec -it rgw2 cephfs-data-scan scan_links
> 2019-11-12 08:46:27.025 7fe775dd7d80 -1 datascan.scan_links: Remove duplicated ino 0x0x100026d17d4 from 0x100013b0d3d/latest.log
> 2019-11-12 08:46:28.665 7fe775dd7d80 -1 datascan.load_table: unable to read mds table 'mds1_inotable': (2) No such file or directory
> 2019-11-12 08:46:28.665 7fe775dd7d80 -1 mds.1.inotable: erasing 0x20000000000 to 0x2000000d665
> 2019-11-12 08:46:28.793 7fe775dd7d80 -1 datascan.load_table: unable to read mds table 'mds2_inotable': (2) No such file or directory
> 2019-11-12 08:46:28.793 7fe775dd7d80 -1 mds.2.inotable: erasing 0x30000000000 to 0x300000228f5
> 2019-11-12 08:46:29.345 7fe775dd7d80 -1 mds.0.snap  updating last_snap 1 -> 3
>
>

please run ceph-mds with debug_mds=20, and send the crash log to me.

Thanks
Yan, Zheng

> >
> >
> > >
> > >
> > >
> > > > - Karsten
> > > >
> > > > -----Original message-----
> > > > From:   Yan, Zheng <ukernel@xxxxxxxxx>
> > > > Sent:   Thu 07-11-2019 14:20
> > > > Subject:        Re:  Re: mds crash loop
> > > > To:     Karsten Nielsen <karsten@xxxxxxxxxx>;
> > > > CC:     ceph-users@xxxxxxx;
> > > > > On Thu, Nov 7, 2019 at 6:40 PM Karsten Nielsen <karsten@xxxxxxxxxx> wrote:
> > > > > >
> > > > > > That is awesome.
> > > > > >
> > > > > > Now I just need to figure out where the lost+found files needs to go.
> > > > > > And what happened to the missing objects for the dirs.
> > > > > >
> > > > >
> > > > > lost+found files are likely files that were deleted. you can keep the
> > > > > lost+found dir for a while, then delete the 'lost+found' directory.
> > > > >
> > > > > for 'missing object' dirs, mv all of them to a temp directory, such as
> > > > > /mnt/cephfs/missing_obj_dirs.
> > > > > Then run command 'ceph daemon mds.x scrub_patch /missing_obj_dirs
> > > > > force recursive repair'. wait a minute, the rm -rf
> > > > > /mnt/cephfs/missing_obj_dirs
> > > > >
> > > > > > Any tool that is able to do that ?
> > > > > >
> > > > > > Thanks
> > > > > > - Karsten
> > > > > >
> > > > > > -----Original message-----
> > > > > > From:   Yan, Zheng <ukernel@xxxxxxxxx>
> > > > > > Sent:   Thu 07-11-2019 09:22
> > > > > > Subject:        Re:  Re: mds crash loop
> > > > > > To:     Karsten Nielsen <karsten@xxxxxxxxxx>;
> > > > > > CC:     ceph-users@xxxxxxx;
> > > > > > > I have tracked down the root cause. See
> > > > > https://tracker.ceph.com/issues/42675
> > > > > > >
> > > > > > > Regards
> > > > > > > Yan, Zheng
> > > > > > >
> > > > > > > On Thu, Nov 7, 2019 at 4:01 PM Karsten Nielsen <karsten@xxxxxxxxxx>
> > > wrote:
> > > > > > > >
> > > > > > > > -----Original message-----
> > > > > > > > From:   Yan, Zheng <ukernel@xxxxxxxxx>
> > > > > > > > Sent:   Thu 07-11-2019 07:21
> > > > > > > > Subject:        Re:  Re: mds crash loop
> > > > > > > > To:     Karsten Nielsen <karsten@xxxxxxxxxx>;
> > > > > > > > CC:     ceph-users@xxxxxxx;
> > > > > > > > > On Thu, Nov 7, 2019 at 5:50 AM Karsten Nielsen
> > <karsten@xxxxxxxxxx>
> > > > > wrote:
> > > > > > > > > >
> > > > > > > > > > -----Original message-----
> > > > > > > > > > From:   Yan, Zheng <ukernel@xxxxxxxxx>
> > > > > > > > > > Sent:   Wed 06-11-2019 14:16
> > > > > > > > > > Subject:        Re:  mds crash loop
> > > > > > > > > > To:     Karsten Nielsen <karsten@xxxxxxxxxx>;
> > > > > > > > > > CC:     ceph-users@xxxxxxx;
> > > > > > > > > > > On Wed, Nov 6, 2019 at 4:42 PM Karsten Nielsen
> > > <karsten@xxxxxxxxxx>
> > > > > > > wrote:
> > > > > > > > > > > >
> > > > > > > > > > > > -----Original message-----
> > > > > > > > > > > > From:   Yan, Zheng <ukernel@xxxxxxxxx>
> > > > > > > > > > > > Sent:   Wed 06-11-2019 08:15
> > > > > > > > > > > > Subject:        Re:  mds crash loop
> > > > > > > > > > > > To:     Karsten Nielsen <karsten@xxxxxxxxxx>;
> > > > > > > > > > > > CC:     ceph-users@xxxxxxx;
> > > > > > > > > > > > > On Tue, Nov 5, 2019 at 5:29 PM Karsten Nielsen
> > > > > <karsten@xxxxxxxxxx>
> > > > > > > > > wrote:
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > Hi,
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > Last week I upgraded my ceph cluster from luminus to
> > mimic
> > > > > 13.2.6
> > > > > > > > > > > > > > It was running fine for a while but yesterday my mds
> > went
> > > > > into a
> > > > > > > crash
> > > > > > > > > > > loop.
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > I have 1 active and 1 standby mds for my cephfs both of
> > > which
> > > > > is
> > > > > > > > > running
> > > > > > > > > > > the
> > > > > > > > > > > > > same crash loop.
> > > > > > > > > > > > > > I am running ceph based on
> > > > > https://hub.docker.com/r/ceph/daemon
> > > > > > > > > version
> > > > > > > > > > > > > v3.2.7-stable-3.2-minic-centos-7-x86_64 with a etcd kv
> > > store.
> > > > > > > > > > > > > >
> > > > > > > > > > > > > > Log details are: https://paste.debian.net/1113943/
> > > > > > > > > > > > > >
> > > > > > > > > > > > >
> > > > > > > > > > > > > please try again with debug_mds=20.  Thanks
> > > > > > > > > > > > >
> > > > > > > > > > > > > Yan, Zheng
> > > > > > > > > > > >
> > > > > > > > > > > > Yes I have set that and had to move to pastebin.com as
> > debian
> > > > > > > apperently
> > > > > > > > > only
> > > > > > > > > > > supports 150k
> > > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > > https://pastebin.com/Gv7c5h54
> > > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > Looks like on-disk root inode is corrupted. have you
> > > encountered any
> > > > > > > > > > > unusually things during the upgrade?
> > > > > > > > > > >
> > > > > > > > > > > please run 'rados -p <cephfs metadata pool> stat
> > > 1.00000000.inode' ,
> > > > > > > > > > > check if the object is modified before or after the 'luminous
> > ->
> > > > > > > > > > > 13.2.6' upgrade.
> > > > > > > > > > > To fix the corrupted object.  Run  'cephfs-data-scan init
> > > > > > > > > > > --force-init'. Then restart mds. After mds become active, run
> > > 'ceph
> > > > > > > > > > > daemon mds.x scrub_path / force repair'
> > > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > I followed the steps I got the mds started but now a lot of
> > files
> > > are
> > > > > in
> > > > > > > > > lost+found 24283 and I have these errors in the mds log
> > > > > > > > > >
> > > > > > > > >  'cephfs-data-scan init --force-init' does not move files into
> > > > > > > > > lost+found. have you ever run other 'cephfs-data-scan foo'
> > command
> > > or
> > > > > > > > > 'cephfs-journal-tool foo' command?
> > > > > > > >
> > > > > > > > I have had a similar problem with the cluster before where I went
> > > through
> > > > > the
> > > > > > > cycle of:
> > > > > > > > https://docs.ceph.com/docs/mimic/cephfs/disaster-recovery/ -> Using
> > an
> > > > > > > alternate metadata pool for recovery
> > > > > > > >
> > > > > > > > I did run the cephfs-journal-tool journal reset command, mostly
> > > because
> > > > > > > cephfs is not that utilized so I thought it was safe to do as after
> > the
> > > > > upgrade
> > > > > > > the cluster has not been used much, so data lose would be minimal -
> > > > > apparently
> > > > > > > I was wrong.
> > > > > > > >
> > > > > > > > >
> > > > > > > > > > 2019-11-06 20:20:18.215 7f0bd9090700  1 mds.0.32011 cluster
> > > recovered.
> > > > > > > > > > 2019-11-06 20:20:19.019 7f0bd2dfa700  0
> > > mds.0.cache.dir(0x100013acfcb)
> > > > > > > > > _fetched missing object for [dir 0x100013acfcb
> > > > > > > /nextcloud/custom_apps/carnet/
> > > > > > > > > [2,head] auth v=0 cv=0/0 ap=1+0+0 state=1073741888|fetching f()
> > n()
> > > > > > > > > hs=0+0,ss=0+0 | waiter=1 authpin=1 0x55d4dc4f5100]
> > > > > > > > > > 2019-11-06 20:20:19.019 7f0bd2dfa700 -1 log_channel(cluster) log
> > > > > [ERR] :
> > > > > > > dir
> > > > > > > > > 0x100013acfcb object missing on disk; some files may be lost
> > > > > > > > > (/nextcloud/custom_apps/carnet)
> > > > > > > > > > 2019-11-06 20:20:19.275 7f0bd2dfa700  0
> > > mds.0.cache.dir(0x100013a3156)
> > > > > > > > > _fetched missing object for [dir 0x100013a3156
> > > > > /nextcloud/custom_apps/mail/
> > > > > > > > > [2,head] auth v=0 cv=0/0 ap=1+0+0 state=1073741888|fetching f()
> > n()
> > > > > > > > > hs=0+0,ss=0+0 | waiter=1 authpin=1 0x55d4dcc40000]
> > > > > > > > > > 2019-11-06 20:20:19.275 7f0bd2dfa700 -1 log_channel(cluster) log
> > > > > [ERR] :
> > > > > > > dir
> > > > > > > > > 0x100013a3156 object missing on disk; some files may be lost
> > > > > > > > > (/nextcloud/custom_apps/mail)
> > > > > > > > > > 2019-11-06 20:20:19.371 7f0bd2dfa700  0
> > > mds.0.cache.dir(0x100013abb3c)
> > > > > > > > > _fetched missing object for [dir 0x100013abb3c
> > > > > > > > > /nextcloud/custom_apps/passwords/ [2,head] auth v=0 cv=0/0
> > ap=1+0+0
> > > > > > > > > state=1073741888|fetching f() n() hs=0+0,ss=0+0 | waiter=1
> > authpin=1
> > > > > > > > > 0x55d4dcc40700]
> > > > > > > > > > 2019-11-06 20:20:19.371 7f0bd2dfa700 -1 log_channel(cluster) log
> > > > > [ERR] :
> > > > > > > dir
> > > > > > > > > 0x100013abb3c object missing on disk; some files may be lost
> > > > > > > > > (/nextcloud/custom_apps/passwords)
> > > > > > > > > > 2019-11-06 20:20:19.383 7f0bd2dfa700  0
> > > mds.0.cache.dir(0x100013a9b9b)
> > > > > > > > > _fetched missing object for [dir 0x100013a9b9b
> > > > > > > > > /nextcloud/custom_apps/phonetrack/ [2,head] auth v=0 cv=0/0
> > ap=1+0+0
> > > > > > > > > state=1073741888|fetching f() n() hs=0+0,ss=0+0 | waiter=1
> > authpin=1
> > > > > > > > > 0x55d4dcc40e00]
> > > > > > > > > > 2019-11-06 20:20:19.383 7f0bd2dfa700 -1 log_channel(cluster) log
> > > > > [ERR] :
> > > > > > > dir
> > > > > > > > > 0x100013a9b9b object missing on disk; some files may be lost
> > > > > > > > > (/nextcloud/custom_apps/phonetrack)
> > > > > > > > > > 2019-11-06 20:20:19.431 7f0bd2dfa700  0
> > > mds.0.cache.dir(0x100013a2659)
> > > > > > > > > _fetched missing object for [dir 0x100013a2659
> > > > > > > > > /nextcloud/custom_apps/richdocuments/ [2,head] auth v=0 cv=0/0
> > > ap=1+0+0
> > > > > > > > > state=1073741888|fetching f() n() hs=0+0,ss=0+0 | waiter=1
> > authpin=1
> > > > > > > > > 0x55d4dcc41500]
> > > > > > > > > > 2019-11-06 20:20:19.431 7f0bd2dfa700 -1 log_channel(cluster) log
> > > > > [ERR] :
> > > > > > > dir
> > > > > > > > > 0x100013a2659 object missing on disk; some files may be lost
> > > > > > > > > (/nextcloud/custom_apps/richdocuments)
> > > > > > > > > > 2019-11-06 20:20:22.360 7f0bd9090700  1 mds.k8s-node-01
> > Updating
> > > MDS
> > > > > map
> > > > > > > to
> > > > > > > > > version 32015 from mon.1
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > > - Karsten
> > > > > > > > > > > >
> > > > > > > > > > > > >
> > > > > > > > > > > > > > Thanks for any hints
> > > > > > > > > > > > > > - Karsten
> > > > > > > > > > > > > > _______________________________________________
> > > > > > > > > > > > > > ceph-users mailing list -- ceph-users@xxxxxxx
> > > > > > > > > > > > > > To unsubscribe send an email to ceph-users-leave@xxxxxxx
> > > > > > > > > > > > >
> > > > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > _______________________________________________
> > > > > > > > > > ceph-users mailing list -- ceph-users@xxxxxxx
> > > > > > > > > > To unsubscribe send an email to ceph-users-leave@xxxxxxx
> > > > > > > > >
> > > > > > > > >
> > > > > > > > _______________________________________________
> > > > > > > > ceph-users mailing list -- ceph-users@xxxxxxx
> > > > > > > > To unsubscribe send an email to ceph-users-leave@xxxxxxx
> > > > > > >
> > > > > > >
> > > > > > _______________________________________________
> > > > > > ceph-users mailing list -- ceph-users@xxxxxxx
> > > > > > To unsubscribe send an email to ceph-users-leave@xxxxxxx
> > > > >
> > > > >
> > >
> > >
> > _______________________________________________
> > ceph-users mailing list -- ceph-users@xxxxxxx
> > To unsubscribe send an email to ceph-users-leave@xxxxxxx
> >
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx




[Index of Archives]     [Information on CEPH]     [Linux Filesystem Development]     [Ceph Development]     [Ceph Large]     [Ceph Dev]     [Linux USB Development]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [xfs]


  Powered by Linux