Re: One OSD always dieing

["Rottmann, Jonas (centron GmbH)" <J.Rottmann@xxxxxxxxxx>]

Hi Andrei,

 

It is the first time I’m running into this. How to fix it? Upgrading with an not fully healthy cluster seams to be not so an great idea.

 

After fixing it I will perform the upgrad ASAP.

 

Thanks for your help so far.

 

Von: Andrei Mikhailovsky [mailto:andrei@xxxxxxxxxx]
Gesendet: Sonntag, 29. Dezember 2013 09:40
An: Rottmann, Jonas (centron GmbH)
Cc: ceph-users@xxxxxxxx
Betreff: Re: [ceph-users] One OSD always dieing

 

 

Jonas,

I've seen this happening on a weekly basis when I was running 0.61 branch as well, however after switching to 0.67 branch it has stopped. Perhaps you should try upgrading

Andrei

 

---

From: "Jonas Rottmann (centron GmbH)" <J.Rottmann@xxxxxxxxxx>
To: "ceph-users@xxxxxxxx" <ceph-users@xxxxxxxx>
Sent: Saturday, 28 December, 2013 9:48:12 AM
Subject: [ceph-users] One OSD always dieing

Hi,

 

One of my OSDs are dieing all the time.  I rebooted one after one every node and assured that all has the same kernel version and glibc.

 

I’m using ceph version 0.61.9 (7440dcd135750839fa0f00263f80722ff6f51e90).

 

Dmesg only shows:

 

[ 5745.366041] init: ceph-osd (ceph/3) main process (2510) killed by ABRT signal

[ 5745.366235] init: ceph-osd (ceph/3) main process ended, respawning

[ 5763.824298] init: ceph-osd (ceph/3) main process (2991) killed by SEGV signal

 

Basically every time this shows up in the logs:

 

2013-12-28 06:35:08.489431 7fc9eccd5700 -1 osd/ReplicatedPG.cc: In function 'ReplicatedPG::RepGather* ReplicatedPG::trim_object(const hobject_t&)' thread 7fc9eccd5700 time 2013-12-28 06:35:08.487862

osd/ReplicatedPG.cc: 1379: FAILED assert(0)

 

If you need more infos I will send them. Please help ! The whole cluster isn’t working proberbly because of this…

_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

 

_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com