Hi!,

I've been migrating my Ovirt storage from iscsi to Gluster. What I've been doing is moving the disks with the VM powered off.

This procedure was fine until I tried to migrate a 100GB disk from a VM with an Oracle DB. During this migration the access to Gluster slowed down, affecting the whole cluster (VDSM timeout in oVirt Engine, VM's entered in unknown state, gluster client disconnected from a brick)

What I found in gluster servers was a XFS error:

XFS: possible memory allocation deadlock in kmem_alloc (mode:0x250)

And this error is due to the number of extents of the file (~ 8M):

gluster01:/gluster/ovirt1_brick_01/brick/89c4e113-1003-4b4e-850e-e7fc5bf2edc6/images/e5b7fc46-4168-4019-a13a-f9b7093d0534# xfs_bmap f9cd555d-fd98-499f-a853-d0ce76eecd37 | wc -l
8627613

I'm using Gluster 3.6.2 and Centos 7 (3.10.0-229.1.2.el7.x86_64 kernel).

Does anyone know why is qemu-img generating this so fragmented file?

Regards.

--
----------------------------------------------------------------
 Adrián Santos Marrero
 Técnico de Sistemas - Área de Infraestructuras TIC
 Servicios de Tecnologías de la Información y Comunicación (STIC)
 Universidad de La Laguna (ULL)
 Teléfono/Phone: +34 922 845089
----------------------------------------------------------------

Este mensaje puede contener información confidencial y/o privilegiada.
Si usted no es el destinatario o lo ha recibido por error debe borrarlo
inmediatamente.
Está estrictamente prohibido por la legislación vigente realizar
cualquier copia, revelación o distribución del contenido de este mensaje
sin autorización expresa.

This e-mail may contain confidential and/or privileged information.
If you are not the intended recipient or have received this e-mail in
error you must destroy it.
Any unauthorised copying, disclosure or distribution of the material in
this e-mail is strictly forbidden by current legislation.