GlusterFS metadata split brain recovery

While investigating an error related to failed documents I came across following error in the GlusterFS healing daemon log file: [2015-11-08 23:22:38.700539] E [afr-self-heal-common.c:197:afr_sh_print_split_brain_log] 0-gfs-volume-vol1-replicate-0: Unable to self-heal contents of ” (possible split-brain). Please delete the file from all but…

GlusterFS internals

GlusterFS stores metadata info in extended attributes which is supported and enabled by default in the XFS file system we use for the bricks. This is different approach then some other distributed storage cluster systems like Ceph for example that…

MySQL High Availability and Load Balancing with Keepalived

MySQL High Availability and Load Balancing with Keepalived

What we want to achieve here is have a MySQL HA two nodes cluster in Master-Master mode and load balance the instances using as less hardware as possible. The role of the LB will be given to Keepalived that will…

VIP(EIP) fail over with Keepalived in Amazon VPC across availability zones

This example covers VIP failover in AWS VPC across AZ’s with Keepalived. The main problem in AWS is that this provider is blocking the multicast traffic in the VPC’s. To circumvent this we need to switch to unicast for the…