Dear ceph users, I just had a problem on a ceph cluster running three nodes with each: - 24 cores - 32Go RAM - 2 SATA disks as OSD - 2 SSD disks using software RAID for sys + journals - libvirt+kvm - ceph network use 2 dedicated Gigabit interfaces with active-passive to two switches Journal partition and ceph osd are using XFS, everything else is running on ext4. OS is Debian Wheezy with kernel 3.10 from backports and ceph is dumpling version. So on one node, both OSDs crashed today with the following logs: - XFS produce regularly those lines on each servers, a bit more on the one that had osd crash and thos lines ocurred just before OSD crash XFS: possible memory allocation deadlock in kmem_alloc (mode:0x250) - ceph osd log is here (crash around 2013-11-28 01:48:33) http://people.easter-eggs.org/~manu/ceph-osd.5.log.1 (15MB) thanks for any hint on solving this issue :) _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |