Created attachment 1473518 [details] logs, core, and screen logs Description of problem: When damage to a disperse volume occurs on a cluster of armhf architecture machines the self heal daemon crashes or consumes large amounts of cpu and can not heal the cluster I am using ubuntu 18.04LTS on Odroid HC-2 hardware (armhf) and have installed gluster 4.1.2 via PPA. I have 12 member nodes each with a single brick. If a cluster node reboots, loses connectivity, or is otherwise unavailable - disperse volumes are unable to heal and the SHD crashes or sometimes consumes large amounts of CPU to no effect once the errant node rejoins the cluster. I have attached an archive containing gluster logs, screen logs and core files. screenlog.sys - Details OS and hardware specs of the system screenlog.gluster - Details cluster specs, volume creation, and volume status before and after a gluster node has become unaavailable and then rejoins the cluster. screenlog.client - The screen log of client mounting and writing files to the volume. Nothing terribly exciting mnt.log.client - Clients mnt.log core - Crash of the shd on a gluster node member gluster.logs - Logs of the crashing glustershd member
glusterfs works fine on ARM from glusterfs-6.x releases. There are minor glitches, but not this basic anymore.